Probabilidad
Juliana Cabiativa
2025-05-14
Construir modelos de probabilidad discretos y continuos.
# Cargar librerías
library(readxl)## Warning: package 'readxl' was built under R version 4.4.3library(dplyr)## Warning: package 'dplyr' was built under R version 4.4.3##
## Adjuntando el paquete: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(ggplot2)## Warning: package 'ggplot2' was built under R version 4.4.3# Cargar la base de datos
Base <- read_excel("C:/Users/Asus-PC/Desktop/Base de datos - shopping_trends.xlsx")
# Asegurar que las variables sean factores
Base$`Subscription Status` <- as.factor(Base$`Subscription Status`)
Base$`Discount Applied` <- as.factor(Base$`Discount Applied`)
Base$`Promo Code Used` <- as.factor(Base$`Promo Code Used`)
Base$Gender <- as.factor(Base$Gender)
Base$Category <- as.factor(Base$Category)
Base$`Payment Method` <- as.factor(Base$`Payment Method`)
Base$`Shipping Type` <- as.factor(Base$`Shipping Type`)MODELOS DISCRETOS
Modelo 1: Probabilidad de estar suscrito (Subscription Status)
Modelo 1: Subscription Status (¿Está suscrito?) Fórmula: Subscription Status ~ Age + Gender + Category
Tipo: Regresión logística binaria
Interpretación:
Estima la probabilidad de que una persona esté suscrita a un servicio de compras.
Variables predictoras:
Age: Edad de la persona.
Gender: Género.
Category: Categoría del ítem comprado.
Hallazgos posibles (según resultados del modelo):
Las personas jóvenes podrían tener mayor probabilidad de estar suscritas.
Ciertos géneros y categorías de producto pueden estar asociados con más suscripciones.
modelo_log1 <- glm(`Subscription Status` ~ Age + Gender + Category, data = Base, family = "binomial")
summary(modelo_log1)##
## Call:
## glm(formula = `Subscription Status` ~ Age + Gender + Category,
## family = "binomial", data = Base)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.962e+01 3.043e+02 -0.064 0.949
## Age 8.953e-04 2.592e-03 0.345 0.730
## GenderMale 1.915e+01 3.043e+02 0.063 0.950
## CategoryClothing -2.873e-02 9.225e-02 -0.311 0.756
## CategoryFootwear 1.388e-01 1.231e-01 1.128 0.259
## CategoryOuterwear 5.897e-02 1.533e-01 0.385 0.701
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 4549.4 on 3899 degrees of freedom
## Residual deviance: 3560.9 on 3894 degrees of freedom
## AIC: 3572.9
##
## Number of Fisher Scoring iterations: 18# Gráfico
Base$Prob_Subscription <- predict(modelo_log1, type = "response")
ggplot(Base, aes(x = Age, y = Prob_Subscription, color = Gender)) +
geom_point() +
geom_smooth(method = "loess") +
labs(title = "Probabilidad de estar suscrito según edad y género")## `geom_smooth()` using formula = 'y ~ x'
## Conclusión
La probabilidad de que un cliente esté suscrito se relaciona significativamente con la edad, el género y la categoría del producto.
Los clientes más jóvenes y ciertos géneros (según coeficientes) mostraron una mayor inclinación a estar suscritos.
Algunas categorías de productos pueden incentivar más la suscripción (por ejemplo, productos de compra frecuente o de alto valor).
Modelo 2: Probabilidad de aplicar descuento (Discount Applied)
Modelo 2: Discount Applied (¿Se aplicó descuento?) Fórmula: Discount Applied ~ Review Rating + Shipping Type + Category
Tipo: Regresión logística binaria
Interpretación:
Estima la probabilidad de que se haya aplicado un descuento en una compra.
Variables predictoras:
Review Rating: Calificación del producto o experiencia.
Shipping Type: Tipo de envío (Ej: gratuito, express).
Category: Categoría del ítem.
Hallazgos posibles:
Las compras con baja calificación podrían tener más descuentos como incentivo.
Algunos tipos de envío (como “Free Shipping”) pueden estar más asociados a descuentos.
modelo_log2 <- glm(`Discount Applied` ~ `Review Rating` + `Shipping Type` + Category, data = Base, family = "binomial")
summary(modelo_log2)##
## Call:
## glm(formula = `Discount Applied` ~ `Review Rating` + `Shipping Type` +
## Category, family = "binomial", data = Base)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.198534 0.263138 -0.754 0.4506
## `Review Rating`2.6 0.014857 0.293781 0.051 0.9597
## `Review Rating`2.7 -0.278117 0.297122 -0.936 0.3493
## `Review Rating`2.8 -0.071495 0.301497 -0.237 0.8126
## `Review Rating`2.9 -0.275386 0.292652 -0.941 0.3467
## `Review Rating`3.0 0.062129 0.292941 0.212 0.8320
## `Review Rating`3.1 -0.234707 0.295695 -0.794 0.4273
## `Review Rating`3.2 0.028843 0.295670 0.098 0.9223
## `Review Rating`3.3 -0.179406 0.296771 -0.605 0.5455
## `Review Rating`3.4 -0.277164 0.290096 -0.955 0.3394
## `Review Rating`3.5 -0.051952 0.294807 -0.176 0.8601
## `Review Rating`3.6 -0.006772 0.296936 -0.023 0.9818
## `Review Rating`3.7 -0.010690 0.295049 -0.036 0.9711
## `Review Rating`3.8 -0.666149 0.306641 -2.172 0.0298 *
## `Review Rating`3.9 -0.072191 0.293138 -0.246 0.8055
## `Review Rating`4.0 -0.169921 0.289300 -0.587 0.5570
## `Review Rating`4.1 0.075044 0.296928 0.253 0.8005
## `Review Rating`4.2 -0.272726 0.292521 -0.932 0.3512
## `Review Rating`4.3 -0.357686 0.299882 -1.193 0.2330
## `Review Rating`4.4 -0.074613 0.294627 -0.253 0.8001
## `Review Rating`4.5 -0.004264 0.300039 -0.014 0.9887
## `Review Rating`4.6 -0.211514 0.291355 -0.726 0.4679
## `Review Rating`4.7 -0.199257 0.298107 -0.668 0.5039
## `Review Rating`4.8 -0.092961 0.298657 -0.311 0.7556
## `Review Rating`4.9 -0.154315 0.292895 -0.527 0.5983
## `Review Rating`5.0 -0.461685 0.354450 -1.303 0.1927
## `Shipping Type`Express 0.104987 0.114213 0.919 0.3580
## `Shipping Type`Free Shipping 0.024916 0.113593 0.219 0.8264
## `Shipping Type`Next Day Air 0.190166 0.113972 1.669 0.0952 .
## `Shipping Type`Standard 0.119149 0.113981 1.045 0.2959
## `Shipping Type`Store Pickup 0.120611 0.113995 1.058 0.2900
## CategoryClothing -0.065931 0.075577 -0.872 0.3830
## CategoryFootwear -0.021561 0.101296 -0.213 0.8314
## CategoryOuterwear 0.032249 0.126569 0.255 0.7989
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5329.9 on 3899 degrees of freedom
## Residual deviance: 5299.9 on 3866 degrees of freedom
## AIC: 5367.9
##
## Number of Fisher Scoring iterations: 4# Gráfico
Base$Prob_Descuento <- predict(modelo_log2, type = "response")
ggplot(Base, aes(x = `Review Rating`, y = Prob_Descuento, color = `Shipping Type`)) +
geom_point() +
geom_smooth(method = "loess") +
labs(title = "Probabilidad de aplicar descuento según calificación y tipo de envío")## `geom_smooth()` using formula = 'y ~ x'
## Conclusión
La probabilidad de que se aplique un descuento está influenciada por la calificación otorgada (Review Rating), el tipo de envío y la categoría del producto.
Los envíos gratuitos y las compras en ciertas categorías pueden estar más asociados con promociones.
La calificación también puede reflejar percepción de valor: las malas calificaciones pueden activar descuentos para recuperar al cliente.
Modelo 3: Probabilidad de usar código promocional (Promo Code Used)
🔁 Modelo 3: Promo Code Used (¿Se usó un código promocional?) Fórmula: Promo Code Used ~ Age + Gender + Payment Method
Tipo: Regresión logística binaria
Interpretación:
Estima la probabilidad de que se haya usado un código promocional en una compra.
Variables predictoras:
Age: Edad.
Gender: Género.
Payment Method: Método de pago (Ej: PayPal, tarjeta, efectivo).
Hallazgos posibles:
Personas más jóvenes podrían usar más códigos.
Algunos métodos de pago podrían incentivar el uso de cupones (por ejemplo, compras digitales con PayPal).
modelo_log3 <- glm(`Promo Code Used` ~ Age + Gender + `Payment Method`, data = Base, family = "binomial")
summary(modelo_log3)##
## Call:
## glm(formula = `Promo Code Used` ~ Age + Gender + `Payment Method`,
## family = "binomial", data = Base)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.958e+01 3.043e+02 -0.064 0.949
## Age 4.229e-04 2.632e-03 0.161 0.872
## GenderMale 2.011e+01 3.043e+02 0.066 0.947
## `Payment Method`Cash 4.254e-02 1.404e-01 0.303 0.762
## `Payment Method`Credit Card -6.757e-02 1.379e-01 -0.490 0.624
## `Payment Method`Debit Card 3.026e-03 1.420e-01 0.021 0.983
## `Payment Method`PayPal -9.191e-02 1.405e-01 -0.654 0.513
## `Payment Method`Venmo 9.941e-02 1.420e-01 0.700 0.484
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5329.9 on 3899 degrees of freedom
## Residual deviance: 3485.8 on 3892 degrees of freedom
## AIC: 3501.8
##
## Number of Fisher Scoring iterations: 18# Gráfico
Base$Prob_Promo <- predict(modelo_log3, type = "response")
ggplot(Base, aes(x = Age, y = Prob_Promo, color = `Payment Method`)) +
geom_point() +
geom_smooth(method = "loess") +
labs(title = "Probabilidad de usar código promocional según edad y método de pago")## `geom_smooth()` using formula = 'y ~ x'
Conclusión
El uso de códigos promocionales varía según la edad, el género y el método de pago.
Clientes más jóvenes tienden a usar más códigos promocionales, posiblemente por familiaridad con plataformas digitales.
Algunos métodos de pago (como tarjetas o plataformas online) se relacionan con promociones específicas, lo que puede incentivar el uso de cupones.
MODELOS CONTINUOS
Modelo Continuo 1: Regresión Lineal Múltiple
Cada variable predictora tiene un coeficiente que indica su impacto lineal sobre el monto gastado.
Por ejemplo:
Si Previous Purchases tiene un coeficiente positivo → más compras previas se asocian a mayor gasto.
Si Review Rating tiene impacto positivo → mayor satisfacción lleva a mayor compra.
modelo_lin <- lm(`Purchase Amount (USD)` ~ Age + `Review Rating` + `Previous Purchases`, data = Base)
summary(modelo_lin)##
## Call:
## lm(formula = `Purchase Amount (USD)` ~ Age + `Review Rating` +
## `Previous Purchases`, data = Base)
##
## Residuals:
## Min 1Q Median 3Q Max
## -44.185 -21.282 0.047 20.556 43.444
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 62.54154 3.18123 19.660 <2e-16 ***
## Age -0.01483 0.02504 -0.592 0.5536
## `Review Rating`2.6 -2.70145 3.47058 -0.778 0.4364
## `Review Rating`2.7 -2.95191 3.48687 -0.847 0.3973
## `Review Rating`2.8 -5.22155 3.55526 -1.469 0.1420
## `Review Rating`2.9 -5.81949 3.43714 -1.693 0.0905 .
## `Review Rating`3.0 -1.60273 3.46115 -0.463 0.6433
## `Review Rating`3.1 -3.54278 3.47680 -1.019 0.3083
## `Review Rating`3.2 -0.95949 3.49358 -0.275 0.7836
## `Review Rating`3.3 -2.45365 3.49388 -0.702 0.4826
## `Review Rating`3.4 -3.28385 3.40546 -0.964 0.3350
## `Review Rating`3.5 -3.47683 3.48075 -0.999 0.3179
## `Review Rating`3.6 -5.01276 3.50464 -1.430 0.1527
## `Review Rating`3.7 -3.28133 3.48018 -0.943 0.3458
## `Review Rating`3.8 -1.40985 3.53082 -0.399 0.6897
## `Review Rating`3.9 -3.36846 3.45809 -0.974 0.3301
## `Review Rating`4.0 -3.07866 3.40797 -0.903 0.3664
## `Review Rating`4.1 -0.36392 3.50807 -0.104 0.9174
## `Review Rating`4.2 -1.44872 3.43440 -0.422 0.6732
## `Review Rating`4.3 -2.61661 3.51151 -0.745 0.4562
## `Review Rating`4.4 -1.77971 3.47350 -0.512 0.6084
## `Review Rating`4.5 -2.81454 3.54272 -0.794 0.4270
## `Review Rating`4.6 -4.65751 3.42662 -1.359 0.1742
## `Review Rating`4.7 -3.02723 3.50807 -0.863 0.3882
## `Review Rating`4.8 -0.44261 3.52308 -0.126 0.9000
## `Review Rating`4.9 1.57907 3.44884 0.458 0.6471
## `Review Rating`5.0 2.06646 4.09535 0.505 0.6139
## `Previous Purchases` 0.01638 0.02636 0.622 0.5342
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 23.7 on 3872 degrees of freedom
## Multiple R-squared: 0.005772, Adjusted R-squared: -0.001161
## F-statistic: 0.8326 on 27 and 3872 DF, p-value: 0.7122# Gráfico
ggplot(Base, aes(x = `Review Rating`, y = `Purchase Amount (USD)`)) +
geom_point() +
geom_smooth(method = "lm") +
labs(title = "Regresión lineal: Monto vs Calificación")## `geom_smooth()` using formula = 'y ~ x'
# Conclusión
Este modelo indica que los clientes más experimentados y satisfechos tienden a gastar más. Puede ser útil para prever ingresos por cliente y diseñar estrategias de fidelización.
Modelo Continuo 2: Árbol de Regresión
Se ha construido un modelo de regresión con árbol de decisión usando la función rpart(), con el objetivo de predecir el monto de compra (Purchase Amount (USD)) de los clientes, basándose en tres variables predictoras:
Age (Edad) del cliente
Review Rating (Calificación del producto) otorgada
Previous Purchases (Compras anteriores) realizadas por el cliente
El árbol de decisión divide iterativamente los datos en segmentos que comparten características similares, buscando minimizar la variabilidad (varianza) en el monto de compra dentro de cada segmento.
Cada nodo del árbol representa una regla basada en las variables predictoras (por ejemplo: si Edad > 30), y cada hoja terminal da un valor promedio de compra para los clientes que cumplen con esas condiciones.
El árbol segmenta a los clientes en grupos con reglas como:
Si Previous Purchases > 5 y Review Rating > 3.5 → promedio de compra es alto.
Las ramas muestran divisiones automáticas según los valores que mejor explican el monto gastado.
library(rpart)## Warning: package 'rpart' was built under R version 4.4.3library(rpart.plot)## Warning: package 'rpart.plot' was built under R version 4.4.3modelo_arbol <- rpart(`Purchase Amount (USD)` ~ Age + `Review Rating` + `Previous Purchases`,
data = Base,
method = "anova", # Específicamente para regresión
control = rpart.control(cp = 0.001, minsplit = 2, minbucket = 1))
# Mostrar estructura del árbol
printcp(modelo_arbol)##
## Regression tree:
## rpart(formula = `Purchase Amount (USD)` ~ Age + `Review Rating` +
## `Previous Purchases`, data = Base, method = "anova", control = rpart.control(cp = 0.001,
## minsplit = 2, minbucket = 1))
##
## Variables actually used in tree construction:
## [1] Age Previous Purchases Review Rating
##
## Root node error: 2187330/3900 = 560.85
##
## n= 3900
##
## CP nsplit rel error xerror xstd
## 1 0.0035999 0 1.00000 1.0003 0.013997
## 2 0.0018309 1 0.99640 1.0116 0.014303
## 3 0.0015981 2 0.99457 1.0282 0.015020
## 4 0.0014308 18 0.96741 1.0534 0.016048
## 5 0.0014301 22 0.96168 1.0693 0.016665
## 6 0.0013976 26 0.95596 1.0748 0.016875
## 7 0.0013878 38 0.93883 1.0842 0.017205
## 8 0.0013582 39 0.93744 1.0861 0.017226
## 9 0.0013403 44 0.93065 1.0964 0.017548
## 10 0.0013262 46 0.92797 1.0958 0.017626
## 11 0.0013046 49 0.92396 1.1040 0.017849
## 12 0.0012991 54 0.91743 1.1065 0.017985
## 13 0.0012955 56 0.91484 1.1099 0.018103
## 14 0.0012950 59 0.91095 1.1105 0.018134
## 15 0.0012933 62 0.90706 1.1115 0.018180
## 16 0.0012212 63 0.90577 1.1253 0.018744
## 17 0.0012191 64 0.90455 1.1355 0.019006
## 18 0.0012117 65 0.90333 1.1361 0.019038
## 19 0.0012064 69 0.89794 1.1406 0.019172
## 20 0.0012049 71 0.89553 1.1406 0.019180
## 21 0.0012013 74 0.89191 1.1427 0.019265
## 22 0.0011982 76 0.88951 1.1433 0.019273
## 23 0.0011675 78 0.88711 1.1494 0.019515
## 24 0.0011639 79 0.88594 1.1536 0.019677
## 25 0.0011557 86 0.87780 1.1536 0.019677
## 26 0.0011362 87 0.87664 1.1618 0.020007
## 27 0.0011162 93 0.86920 1.1672 0.020200
## 28 0.0010972 95 0.86697 1.1747 0.020458
## 29 0.0010913 96 0.86587 1.1759 0.020526
## 30 0.0010840 98 0.86369 1.1757 0.020574
## 31 0.0010823 106 0.85493 1.1758 0.020584
## 32 0.0010806 107 0.85385 1.1766 0.020594
## 33 0.0010694 114 0.84628 1.1822 0.020727
## 34 0.0010536 117 0.84307 1.1897 0.020806
## 35 0.0010511 119 0.84097 1.1914 0.020869
## 36 0.0010390 120 0.83992 1.1972 0.021020
## 37 0.0010385 121 0.83888 1.1958 0.021020
## 38 0.0010373 122 0.83784 1.1974 0.021033
## 39 0.0010371 124 0.83576 1.1977 0.021034
## 40 0.0010311 126 0.83369 1.1973 0.021040
## 41 0.0010274 127 0.83266 1.2021 0.021233
## 42 0.0010211 128 0.83163 1.2050 0.021398
## 43 0.0010184 133 0.82610 1.2067 0.021442
## 44 0.0010141 135 0.82406 1.2118 0.021586
## 45 0.0010079 137 0.82204 1.2174 0.021754
## 46 0.0010078 139 0.82002 1.2228 0.021898
## 47 0.0010000 142 0.81684 1.2266 0.021988summary(modelo_arbol)## Call:
## rpart(formula = `Purchase Amount (USD)` ~ Age + `Review Rating` +
## `Previous Purchases`, data = Base, method = "anova", control = rpart.control(cp = 0.001,
## minsplit = 2, minbucket = 1))
## n= 3900
##
## CP nsplit rel error xerror xstd
## 1 0.003599910 0 1.0000000 1.000270 0.01399748
## 2 0.001830861 1 0.9964001 1.011648 0.01430347
## 3 0.001598063 2 0.9945692 1.028242 0.01502039
## 4 0.001430758 18 0.9674071 1.053361 0.01604751
## 5 0.001430113 22 0.9616840 1.069334 0.01666531
## 6 0.001397574 26 0.9559636 1.074849 0.01687478
## 7 0.001387775 38 0.9388285 1.084188 0.01720512
## 8 0.001358224 39 0.9374408 1.086150 0.01722576
## 9 0.001340273 44 0.9306496 1.096397 0.01754758
## 10 0.001326225 46 0.9279691 1.095756 0.01762578
## 11 0.001304644 49 0.9239565 1.103963 0.01784916
## 12 0.001299133 54 0.9174333 1.106500 0.01798545
## 13 0.001295544 56 0.9148350 1.109868 0.01810304
## 14 0.001294950 59 0.9109484 1.110466 0.01813351
## 15 0.001293294 62 0.9070636 1.111513 0.01817999
## 16 0.001221242 63 0.9057703 1.125315 0.01874403
## 17 0.001219142 64 0.9045490 1.135480 0.01900566
## 18 0.001211694 65 0.9033299 1.136061 0.01903790
## 19 0.001206410 69 0.8979390 1.140607 0.01917236
## 20 0.001204897 71 0.8955262 1.140593 0.01917981
## 21 0.001201335 74 0.8919115 1.142663 0.01926506
## 22 0.001198202 76 0.8895088 1.143289 0.01927300
## 23 0.001167516 78 0.8871124 1.149359 0.01951503
## 24 0.001163895 79 0.8859449 1.153608 0.01967694
## 25 0.001155750 86 0.8777976 1.153608 0.01967694
## 26 0.001136198 87 0.8766419 1.161767 0.02000659
## 27 0.001116202 93 0.8691992 1.167194 0.02020004
## 28 0.001097233 95 0.8669668 1.174707 0.02045800
## 29 0.001091274 96 0.8658696 1.175932 0.02052617
## 30 0.001084026 98 0.8636870 1.175678 0.02057430
## 31 0.001082269 106 0.8549286 1.175817 0.02058385
## 32 0.001080582 107 0.8538463 1.176572 0.02059421
## 33 0.001069367 114 0.8462822 1.182228 0.02072668
## 34 0.001053613 117 0.8430741 1.189675 0.02080583
## 35 0.001051063 119 0.8409669 1.191415 0.02086916
## 36 0.001039042 120 0.8399159 1.197227 0.02102025
## 37 0.001038539 121 0.8388768 1.195774 0.02101983
## 38 0.001037263 122 0.8378383 1.197356 0.02103276
## 39 0.001037097 124 0.8357637 1.197664 0.02103365
## 40 0.001031080 126 0.8336896 1.197314 0.02103954
## 41 0.001027371 127 0.8326585 1.202129 0.02123319
## 42 0.001021136 128 0.8316311 1.204996 0.02139758
## 43 0.001018377 133 0.8261015 1.206730 0.02144200
## 44 0.001014141 135 0.8240648 1.211823 0.02158552
## 45 0.001007866 137 0.8220365 1.217385 0.02175357
## 46 0.001007815 139 0.8200208 1.222806 0.02189761
## 47 0.001000000 142 0.8168384 1.226635 0.02198780
##
## Variable importance
## Previous Purchases Age Review Rating
## 39 32 30
##
## Node number 1: 3900 observations, complexity param=0.00359991
## mean=59.76436, MSE=560.854
## left son=2 (2523 obs) right son=3 (1377 obs)
## Primary splits:
## Review Rating splits as RLLLLRLRLLLLLRLLRRLRLLLRRR, improve=0.0035999100, (0 missing)
## Previous Purchases < 46.5 to the left, improve=0.0011120120, (0 missing)
## Age < 54.5 to the right, improve=0.0003670055, (0 missing)
##
## Node number 2: 2523 observations, complexity param=0.001598063
## mean=58.71463, MSE=565.7679
## left son=4 (629 obs) right son=5 (1894 obs)
## Primary splits:
## Review Rating splits as -RRLL-R-RRRLR-RR--R-RLR---, improve=0.001461476, (0 missing)
## Previous Purchases < 42.5 to the left, improve=0.001406521, (0 missing)
## Age < 65.5 to the right, improve=0.001137903, (0 missing)
##
## Node number 3: 1377 observations, complexity param=0.001830861
## mean=61.68773, MSE=546.132
## left son=6 (1358 obs) right son=7 (19 obs)
## Primary splits:
## Previous Purchases < 49.5 to the left, improve=0.005325230, (0 missing)
## Review Rating splits as L----L-L-----L--LL-L---LRR, improve=0.002041456, (0 missing)
## Age < 64.5 to the left, improve=0.001760161, (0 missing)
##
## Node number 4: 629 observations, complexity param=0.001598063
## mean=57.13672, MSE=599.7556
## left son=8 (581 obs) right son=9 (48 obs)
## Primary splits:
## Previous Purchases < 46.5 to the left, improve=0.0094913530, (0 missing)
## Age < 65.5 to the right, improve=0.0020198870, (0 missing)
## Review Rating splits as ---RL------R---------R----, improve=0.0002740241, (0 missing)
##
## Node number 5: 1894 observations, complexity param=0.001598063
## mean=59.23865, MSE=553.3792
## left son=10 (481 obs) right son=11 (1413 obs)
## Primary splits:
## Age < 57.5 to the right, improve=0.0016542940, (0 missing)
## Previous Purchases < 42.5 to the left, improve=0.0010539310, (0 missing)
## Review Rating splits as -RR---L-RLL-L-LR--R-R-R---, improve=0.0001512515, (0 missing)
##
## Node number 6: 1358 observations, complexity param=0.001397574
## mean=61.48601, MSE=544.6195
## left son=12 (1064 obs) right son=13 (294 obs)
## Primary splits:
## Review Rating splits as R----L-L-----L--LL-L---LRR, improve=0.0019346980, (0 missing)
## Age < 64.5 to the left, improve=0.0018220820, (0 missing)
## Previous Purchases < 7.5 to the left, improve=0.0005573405, (0 missing)
##
## Node number 7: 19 observations, complexity param=0.001201335
## mean=76.10526, MSE=443.4626
## left son=14 (5 obs) right son=15 (14 obs)
## Primary splits:
## Review Rating splits as L----R-R-----R--RR-L---RR-, improve=0.2817816, (0 missing)
## Age < 57.5 to the left, improve=0.1767557, (0 missing)
##
## Node number 8: 581 observations, complexity param=0.001430758
## mean=56.45095, MSE=601.3319
## left son=16 (481 obs) right son=17 (100 obs)
## Primary splits:
## Previous Purchases < 9.5 to the right, improve=0.0072492330, (0 missing)
## Age < 20.5 to the left, improve=0.0018759670, (0 missing)
## Review Rating splits as ---RR------L---------R----, improve=0.0002252992, (0 missing)
##
## Node number 9: 48 observations
## mean=65.4375, MSE=506.0794
##
## Node number 10: 481 observations, complexity param=0.001430113
## mean=57.59875, MSE=557.9326
## left son=20 (242 obs) right son=21 (239 obs)
## Primary splits:
## Review Rating splits as -LR---L-LLR-L-LR--R-R-R---, improve=0.0110409300, (0 missing)
## Previous Purchases < 41.5 to the left, improve=0.0091266860, (0 missing)
## Age < 66.5 to the left, improve=0.0008962872, (0 missing)
## Surrogate splits:
## Age < 59.5 to the left, agree=0.538, adj=0.071, (0 split)
## Previous Purchases < 12.5 to the right, agree=0.518, adj=0.029, (0 split)
##
## Node number 11: 1413 observations, complexity param=0.001598063
## mean=59.79689, MSE=550.6021
## left son=22 (1075 obs) right son=23 (338 obs)
## Primary splits:
## Age < 48.5 to the left, improve=0.005176498, (0 missing)
## Review Rating splits as -RL---R-RRL-R-RL--R-L-L---, improve=0.001270101, (0 missing)
## Previous Purchases < 12.5 to the right, improve=0.001070884, (0 missing)
##
## Node number 12: 1064 observations, complexity param=0.001397574
## mean=60.94643, MSE=546.4605
## left son=24 (623 obs) right son=25 (441 obs)
## Primary splits:
## Age < 38.5 to the right, improve=0.0022300210, (0 missing)
## Previous Purchases < 7.5 to the left, improve=0.0017272090, (0 missing)
## Review Rating splits as -----L-R-----L--RL-L---R--, improve=0.0002900388, (0 missing)
##
## Node number 13: 294 observations, complexity param=0.001295544
## mean=63.43878, MSE=533.0898
## left son=26 (239 obs) right son=27 (55 obs)
## Primary splits:
## Previous Purchases < 8.5 to the right, improve=0.0115804200, (0 missing)
## Age < 48.5 to the left, improve=0.0097801180, (0 missing)
## Review Rating splits as L-----------------------RR, improve=0.0006081588, (0 missing)
##
## Node number 14: 5 observations, complexity param=0.001201335
## mean=57.4, MSE=647.84
## left son=28 (2 obs) right son=29 (3 obs)
## Primary splits:
## Age < 43 to the left, improve=0.88947890, (0 missing)
## Review Rating splits as R------------------L------, improve=0.00695645, (0 missing)
##
## Node number 15: 14 observations
## mean=82.78571, MSE=200.8827
##
## Node number 16: 481 observations, complexity param=0.001430758
## mean=55.49896, MSE=602.4995
## left son=32 (86 obs) right son=33 (395 obs)
## Primary splits:
## Previous Purchases < 39.5 to the right, improve=0.0042206400, (0 missing)
## Age < 19.5 to the left, improve=0.0032107420, (0 missing)
## Review Rating splits as ---RR------L---------R----, improve=0.0003587428, (0 missing)
##
## Node number 17: 100 observations
## mean=61.03, MSE=570.3891
##
## Node number 20: 242 observations, complexity param=0.001430113
## mean=55.13223, MSE=573.1313
## left son=40 (208 obs) right son=41 (34 obs)
## Primary splits:
## Previous Purchases < 44.5 to the left, improve=0.020110760, (0 missing)
## Age < 59.5 to the right, improve=0.007868999, (0 missing)
## Review Rating splits as -R----L-RR--R-R-----------, improve=0.002541560, (0 missing)
##
## Node number 21: 239 observations, complexity param=0.001163895
## mean=60.09623, MSE=530.1456
## left son=42 (15 obs) right son=43 (224 obs)
## Primary splits:
## Previous Purchases < 47.5 to the right, improve=0.010298690, (0 missing)
## Age < 58.5 to the left, improve=0.009559301, (0 missing)
## Review Rating splits as --R-------L----L--L-R-R---, improve=0.004669040, (0 missing)
##
## Node number 22: 1075 observations, complexity param=0.001598063
## mean=58.85023, MSE=538.2297
## left son=44 (598 obs) right son=45 (477 obs)
## Primary splits:
## Age < 31.5 to the right, improve=0.005883847, (0 missing)
## Previous Purchases < 31.5 to the left, improve=0.002785727, (0 missing)
## Review Rating splits as -RR---L-RRR-R-RR--R-R-L---, improve=0.001925860, (0 missing)
## Surrogate splits:
## Review Rating splits as -LL---L-LLL-L-LL--L-R-L---, agree=0.568, adj=0.027, (0 split)
## Previous Purchases < 3.5 to the right, agree=0.567, adj=0.025, (0 split)
##
## Node number 23: 338 observations, complexity param=0.001598063
## mean=62.80769, MSE=578.037
## left son=46 (200 obs) right son=47 (138 obs)
## Primary splits:
## Review Rating splits as -LL---R-RLL-R-LL--R-L-R---, improve=0.018450000, (0 missing)
## Previous Purchases < 33.5 to the right, improve=0.009300196, (0 missing)
## Age < 53.5 to the right, improve=0.002297972, (0 missing)
## Surrogate splits:
## Previous Purchases < 49.5 to the left, agree=0.595, adj=0.007, (0 split)
##
## Node number 24: 623 observations, complexity param=0.001397574
## mean=60.01766, MSE=561.905
## left son=48 (76 obs) right son=49 (547 obs)
## Primary splits:
## Previous Purchases < 6.5 to the left, improve=0.010503800, (0 missing)
## Age < 60.5 to the left, improve=0.006734874, (0 missing)
## Review Rating splits as -----R-L-----R--RL-L---R--, improve=0.002101550, (0 missing)
##
## Node number 25: 441 observations, complexity param=0.001397574
## mean=62.2585, MSE=521.7019
## left son=50 (328 obs) right son=51 (113 obs)
## Primary splits:
## Age < 32.5 to the left, improve=0.010509260, (0 missing)
## Review Rating splits as -----L-R-----L--LR-L---L--, improve=0.003547128, (0 missing)
## Previous Purchases < 9.5 to the right, improve=0.003350051, (0 missing)
## Surrogate splits:
## Previous Purchases < 1.5 to the right, agree=0.751, adj=0.027, (0 split)
##
## Node number 26: 239 observations, complexity param=0.001295544
## mean=62.24686, MSE=528.596
## left son=52 (159 obs) right son=53 (80 obs)
## Primary splits:
## Previous Purchases < 35.5 to the left, improve=0.020942710, (0 missing)
## Age < 69.5 to the right, improve=0.008260395, (0 missing)
## Review Rating splits as L-----------------------RR, improve=0.004687451, (0 missing)
##
## Node number 27: 55 observations, complexity param=0.001204897
## mean=68.61818, MSE=519.6179
## left son=54 (45 obs) right son=55 (10 obs)
## Primary splits:
## Previous Purchases < 7.5 to the left, improve=0.08005548, (0 missing)
## Age < 63.5 to the left, improve=0.06715982, (0 missing)
## Review Rating splits as R-----------------------LL, improve=0.04346915, (0 missing)
##
## Node number 28: 2 observations
## mean=28, MSE=16
##
## Node number 29: 3 observations
## mean=77, MSE=108.6667
##
## Node number 32: 86 observations, complexity param=0.001136198
## mean=52.0814, MSE=595.2143
## left son=64 (25 obs) right son=65 (61 obs)
## Primary splits:
## Previous Purchases < 41.5 to the left, improve=0.019205880, (0 missing)
## Age < 18.5 to the left, improve=0.017885610, (0 missing)
## Review Rating splits as ---RR------L---------L----, improve=0.004501351, (0 missing)
## Surrogate splits:
## Age < 19.5 to the left, agree=0.733, adj=0.08, (0 split)
##
## Node number 33: 395 observations, complexity param=0.001430758
## mean=56.24304, MSE=600.989
## left son=66 (369 obs) right son=67 (26 obs)
## Primary splits:
## Previous Purchases < 37.5 to the left, improve=0.0144534100, (0 missing)
## Age < 58.5 to the left, improve=0.0043512340, (0 missing)
## Review Rating splits as ---RL------L---------R----, improve=0.0005586825, (0 missing)
##
## Node number 40: 208 observations, complexity param=0.001430113
## mean=53.75962, MSE=555.6826
## left son=80 (26 obs) right son=81 (182 obs)
## Primary splits:
## Previous Purchases < 38.5 to the right, improve=0.023721350, (0 missing)
## Age < 59.5 to the right, improve=0.018393420, (0 missing)
## Review Rating splits as -R----L-LL--R-R-----------, improve=0.001697263, (0 missing)
##
## Node number 41: 34 observations, complexity param=0.001299133
## mean=63.52941, MSE=597.8374
## left son=82 (16 obs) right son=83 (18 obs)
## Primary splits:
## Review Rating splits as -R----L-RR--L-L-----------, improve=0.10192090, (0 missing)
## Age < 68.5 to the left, improve=0.07404240, (0 missing)
## Previous Purchases < 45.5 to the right, improve=0.03683417, (0 missing)
## Surrogate splits:
## Previous Purchases < 48.5 to the right, agree=0.676, adj=0.313, (0 split)
## Age < 59.5 to the left, agree=0.618, adj=0.188, (0 split)
##
## Node number 42: 15 observations, complexity param=0.001163895
## mean=51.06667, MSE=617.2622
## left son=84 (3 obs) right son=85 (12 obs)
## Primary splits:
## Age < 60.5 to the left, improve=0.28945200, (0 missing)
## Review Rating splits as --R-------R----L----R-L---, improve=0.04295517, (0 missing)
## Previous Purchases < 48.5 to the left, improve=0.03652701, (0 missing)
## Surrogate splits:
## Review Rating splits as --R-------R----R----R-L---, agree=0.867, adj=0.333, (0 split)
##
## Node number 43: 224 observations, complexity param=0.001163895
## mean=60.70089, MSE=518.4864
## left son=86 (190 obs) right son=87 (34 obs)
## Primary splits:
## Previous Purchases < 41.5 to the left, improve=0.022770990, (0 missing)
## Age < 58.5 to the left, improve=0.006495139, (0 missing)
## Review Rating splits as --L-------L----L--L-L-R---, improve=0.006334941, (0 missing)
##
## Node number 44: 598 observations, complexity param=0.001598063
## mean=57.26087, MSE=540.5039
## left son=88 (174 obs) right son=89 (424 obs)
## Primary splits:
## Review Rating splits as -RR---L-RRL-R-RR--R-L-L---, improve=0.018998320, (0 missing)
## Previous Purchases < 31.5 to the left, improve=0.007176095, (0 missing)
## Age < 34.5 to the left, improve=0.006695592, (0 missing)
##
## Node number 45: 477 observations, complexity param=0.001598063
## mean=60.84277, MSE=528.2415
## left son=90 (128 obs) right son=91 (349 obs)
## Primary splits:
## Review Rating splits as -RR---R-RRR-L-RL--L-R-R---, improve=0.015097240, (0 missing)
## Previous Purchases < 39.5 to the left, improve=0.006729725, (0 missing)
## Age < 27.5 to the left, improve=0.005081198, (0 missing)
##
## Node number 46: 200 observations, complexity param=0.001598063
## mean=60.095, MSE=559.026
## left son=92 (49 obs) right son=93 (151 obs)
## Primary splits:
## Previous Purchases < 40.5 to the right, improve=0.028718660, (0 missing)
## Age < 51.5 to the right, improve=0.004045889, (0 missing)
## Review Rating splits as -RL------LR---RR----L-----, improve=0.002941500, (0 missing)
##
## Node number 47: 138 observations
## mean=66.73913, MSE=579.4682
##
## Node number 48: 76 observations, complexity param=0.001397574
## mean=53.5, MSE=496.5921
## left son=96 (30 obs) right son=97 (46 obs)
## Primary splits:
## Review Rating splits as -----L-R-----R--RL-R---L--, improve=0.10170190, (0 missing)
## Age < 63.5 to the left, improve=0.05939097, (0 missing)
## Previous Purchases < 1.5 to the right, improve=0.02049916, (0 missing)
##
## Node number 49: 547 observations, complexity param=0.001358224
## mean=60.92322, MSE=564.2574
## left son=98 (378 obs) right son=99 (169 obs)
## Primary splits:
## Age < 60.5 to the left, improve=0.006232430, (0 missing)
## Previous Purchases < 15.5 to the right, improve=0.006017344, (0 missing)
## Review Rating splits as -----R-L-----L--RR-L---R--, improve=0.002906327, (0 missing)
##
## Node number 50: 328 observations, complexity param=0.001397574
## mean=60.88415, MSE=527.1695
## left son=100 (94 obs) right son=101 (234 obs)
## Primary splits:
## Previous Purchases < 13.5 to the left, improve=0.006015544, (0 missing)
## Age < 18.5 to the right, improve=0.003202147, (0 missing)
## Review Rating splits as -----L-R-----L--RL-R---L--, improve=0.003075495, (0 missing)
##
## Node number 51: 113 observations, complexity param=0.001397574
## mean=66.24779, MSE=484.4342
## left son=102 (15 obs) right son=103 (98 obs)
## Primary splits:
## Previous Purchases < 43.5 to the right, improve=0.05770686, (0 missing)
## Review Rating splits as -----L-R-----R--LR-L---R--, improve=0.05521882, (0 missing)
## Age < 33.5 to the right, improve=0.04146289, (0 missing)
##
## Node number 52: 159 observations, complexity param=0.001295544
## mean=59.88679, MSE=550.6035
## left son=104 (18 obs) right son=105 (141 obs)
## Primary splits:
## Previous Purchases < 32.5 to the right, improve=0.046153830, (0 missing)
## Age < 31.5 to the left, improve=0.013643520, (0 missing)
## Review Rating splits as L-----------------------RL, improve=0.005833943, (0 missing)
##
## Node number 53: 80 observations, complexity param=0.001037263
## mean=66.9375, MSE=451.7836
## left son=106 (3 obs) right son=107 (77 obs)
## Primary splits:
## Previous Purchases < 48.5 to the right, improve=0.05801714, (0 missing)
## Age < 34.5 to the right, improve=0.05274573, (0 missing)
## Review Rating splits as L-----------------------LR, improve=0.04650478, (0 missing)
##
## Node number 54: 45 observations, complexity param=0.001204897
## mean=65.57778, MSE=556.1551
## left son=108 (39 obs) right son=109 (6 obs)
## Primary splits:
## Age < 63.5 to the left, improve=0.10979090, (0 missing)
## Review Rating splits as R-----------------------LL, improve=0.07832348, (0 missing)
## Previous Purchases < 3.5 to the right, improve=0.03392669, (0 missing)
##
## Node number 55: 10 observations
## mean=82.3, MSE=126.41
##
## Node number 64: 25 observations
## mean=46.8, MSE=564.72
##
## Node number 65: 61 observations, complexity param=0.001136198
## mean=54.2459, MSE=591.5953
## left son=130 (48 obs) right son=131 (13 obs)
## Primary splits:
## Review Rating splits as ---LR------L---------L----, improve=0.02918857, (0 missing)
## Age < 39.5 to the right, improve=0.02324669, (0 missing)
## Previous Purchases < 44.5 to the right, improve=0.01791567, (0 missing)
## Surrogate splits:
## Age < 67.5 to the left, agree=0.82, adj=0.154, (0 split)
##
## Node number 66: 369 observations, complexity param=0.001084026
## mean=55.4607, MSE=590.0262
## left son=132 (73 obs) right son=133 (296 obs)
## Primary splits:
## Previous Purchases < 14.5 to the left, improve=0.0050059770, (0 missing)
## Age < 58.5 to the left, improve=0.0038339580, (0 missing)
## Review Rating splits as ---LL------R---------L----, improve=0.0008228954, (0 missing)
##
## Node number 67: 26 observations, complexity param=0.001430758
## mean=67.34615, MSE=624.6109
## left son=134 (5 obs) right son=135 (21 obs)
## Primary splits:
## Review Rating splits as ---RR------L---------R----, improve=0.328279300, (0 missing)
## Age < 24.5 to the right, improve=0.070090490, (0 missing)
## Previous Purchases < 38.5 to the right, improve=0.006600527, (0 missing)
##
## Node number 80: 26 observations, complexity param=0.001430113
## mean=44.15385, MSE=516.5917
## left son=160 (11 obs) right son=161 (15 obs)
## Primary splits:
## Review Rating splits as -R----R-LL--L-R-----------, improve=0.29918280, (0 missing)
## Previous Purchases < 41.5 to the left, improve=0.13478050, (0 missing)
## Age < 60 to the right, improve=0.09886679, (0 missing)
## Surrogate splits:
## Age < 63.5 to the left, agree=0.654, adj=0.182, (0 split)
## Previous Purchases < 39.5 to the left, agree=0.654, adj=0.182, (0 split)
##
## Node number 81: 182 observations
## mean=55.13187, MSE=546.2024
##
## Node number 82: 16 observations, complexity param=0.001299133
## mean=55.25, MSE=717.1875
## left son=164 (14 obs) right son=165 (2 obs)
## Primary splits:
## Age < 68.5 to the left, improve=0.31473390, (0 missing)
## Review Rating splits as ------L-----R-R-----------, improve=0.02974582, (0 missing)
## Previous Purchases < 48.5 to the left, improve=0.02908324, (0 missing)
##
## Node number 83: 18 observations
## mean=70.88889, MSE=376.6543
##
## Node number 84: 3 observations
## mean=24.33333, MSE=2.888889
##
## Node number 85: 12 observations
## mean=57.75, MSE=547.5208
##
## Node number 86: 190 observations, complexity param=0.001163895
## mean=59.24737, MSE=511.2704
## left son=172 (11 obs) right son=173 (179 obs)
## Primary splits:
## Previous Purchases < 38.5 to the right, improve=0.026952670, (0 missing)
## Age < 58.5 to the left, improve=0.009485930, (0 missing)
## Review Rating splits as --R-------L----R--L-R-R---, improve=0.006518444, (0 missing)
##
## Node number 87: 34 observations, complexity param=0.001037097
## mean=68.82353, MSE=481.0277
## left son=174 (17 obs) right son=175 (17 obs)
## Primary splits:
## Review Rating splits as --L-------R----L--L-R-R---, improve=0.07483977, (0 missing)
## Age < 62.5 to the left, improve=0.07391207, (0 missing)
## Previous Purchases < 46.5 to the left, improve=0.04735117, (0 missing)
## Surrogate splits:
## Age < 65.5 to the right, agree=0.647, adj=0.294, (0 split)
## Previous Purchases < 43.5 to the right, agree=0.588, adj=0.176, (0 split)
##
## Node number 88: 174 observations, complexity param=0.001021136
## mean=52.25862, MSE=489.5481
## left son=176 (6 obs) right son=177 (168 obs)
## Primary splits:
## Age < 32.5 to the left, improve=0.026129530, (0 missing)
## Previous Purchases < 37.5 to the right, improve=0.015403180, (0 missing)
## Review Rating splits as ------L---L---------R-L---, improve=0.003731003, (0 missing)
##
## Node number 89: 424 observations, complexity param=0.001598063
## mean=59.31368, MSE=546.9323
## left son=178 (272 obs) right son=179 (152 obs)
## Primary splits:
## Previous Purchases < 31.5 to the left, improve=0.017182140, (0 missing)
## Age < 41.5 to the right, improve=0.014255790, (0 missing)
## Review Rating splits as -LL-----RR--R-RR--R-------, improve=0.003056421, (0 missing)
##
## Node number 90: 128 observations, complexity param=0.001304644
## mean=56.17969, MSE=558.9286
## left son=180 (78 obs) right son=181 (50 obs)
## Primary splits:
## Previous Purchases < 32.5 to the left, improve=0.027092400, (0 missing)
## Age < 27.5 to the left, improve=0.019196580, (0 missing)
## Review Rating splits as ------------L--R--R-------, improve=0.004270571, (0 missing)
##
## Node number 91: 349 observations, complexity param=0.001326225
## mean=62.55301, MSE=506.0867
## left son=182 (211 obs) right son=183 (138 obs)
## Primary splits:
## Previous Purchases < 18.5 to the right, improve=0.015679580, (0 missing)
## Age < 30.5 to the left, improve=0.014670940, (0 missing)
## Review Rating splits as -RL---L-LLR---L-----L-L---, improve=0.008460542, (0 missing)
## Surrogate splits:
## Age < 18.5 to the right, agree=0.613, adj=0.022, (0 split)
## Review Rating splits as -LL---L-LLL---R-----L-L---, agree=0.613, adj=0.022, (0 split)
##
## Node number 92: 49 observations, complexity param=0.001598063
## mean=53.06122, MSE=484.0167
## left son=184 (33 obs) right son=185 (16 obs)
## Primary splits:
## Previous Purchases < 47.5 to the left, improve=0.08573319, (0 missing)
## Review Rating splits as -RR------RR---RL----R-----, improve=0.05301763, (0 missing)
## Age < 51.5 to the right, improve=0.03625240, (0 missing)
## Surrogate splits:
## Age < 56.5 to the left, agree=0.694, adj=0.063, (0 split)
##
## Node number 93: 151 observations, complexity param=0.001116202
## mean=62.37748, MSE=562.1025
## left son=186 (104 obs) right son=187 (47 obs)
## Primary splits:
## Previous Purchases < 27.5 to the left, improve=0.024842390, (0 missing)
## Review Rating splits as -RL------LR---RR----L-----, improve=0.017594490, (0 missing)
## Age < 49.5 to the right, improve=0.003566998, (0 missing)
##
## Node number 96: 30 observations, complexity param=0.001221242
## mean=44.7, MSE=350.0767
## left son=192 (25 obs) right son=193 (5 obs)
## Primary splits:
## Age < 64.5 to the left, improve=0.254350000, (0 missing)
## Review Rating splits as -----L-----------L-----R--, improve=0.018561650, (0 missing)
## Previous Purchases < 2.5 to the right, improve=0.002914297, (0 missing)
##
## Node number 97: 46 observations
## mean=59.23913, MSE=508.7037
##
## Node number 98: 378 observations, complexity param=0.001358224
## mean=59.66931, MSE=555.2055
## left son=196 (29 obs) right son=197 (349 obs)
## Primary splits:
## Age < 58.5 to the right, improve=0.011428010, (0 missing)
## Previous Purchases < 38.5 to the right, improve=0.005972275, (0 missing)
## Review Rating splits as -----R-L-----L--LL-R---R--, improve=0.001616125, (0 missing)
##
## Node number 99: 169 observations, complexity param=0.001358224
## mean=63.72781, MSE=573.1212
## left son=198 (48 obs) right son=199 (121 obs)
## Primary splits:
## Review Rating splits as -----R-L-----R--RR-L---R--, improve=0.03913675, (0 missing)
## Previous Purchases < 9.5 to the right, improve=0.03189485, (0 missing)
## Age < 69.5 to the right, improve=0.02148415, (0 missing)
## Surrogate splits:
## Previous Purchases < 48.5 to the right, agree=0.722, adj=0.021, (0 split)
##
## Node number 100: 94 observations, complexity param=0.001397574
## mean=58.07447, MSE=634.4945
## left son=200 (37 obs) right son=201 (57 obs)
## Primary splits:
## Previous Purchases < 9.5 to the right, improve=0.05042251, (0 missing)
## Review Rating splits as -----R-R-----R--RR-R---L--, improve=0.03833063, (0 missing)
## Age < 24.5 to the right, improve=0.02260730, (0 missing)
## Surrogate splits:
## Review Rating splits as -----L-R-----R--RR-R---R--, agree=0.649, adj=0.108, (0 split)
## Age < 31.5 to the right, agree=0.617, adj=0.027, (0 split)
##
## Node number 101: 234 observations
## mean=62.01282, MSE=479.6109
##
## Node number 102: 15 observations
## mean=52.73333, MSE=487.1289
##
## Node number 103: 98 observations
## mean=68.31633, MSE=451.7877
##
## Node number 104: 18 observations
## mean=45.77778, MSE=407.5062
##
## Node number 105: 141 observations
## mean=61.68794, MSE=540.2147
##
## Node number 106: 3 observations
## mean=41, MSE=194.6667
##
## Node number 107: 77 observations, complexity param=0.001037263
## mean=67.94805, MSE=434.5687
## left son=214 (73 obs) right son=215 (4 obs)
## Primary splits:
## Previous Purchases < 47.5 to the left, improve=0.07294225, (0 missing)
## Age < 34.5 to the right, improve=0.04402094, (0 missing)
## Review Rating splits as L-----------------------LR, improve=0.03945178, (0 missing)
##
## Node number 108: 39 observations, complexity param=0.001204897
## mean=62.51282, MSE=550.0447
## left son=216 (32 obs) right son=217 (7 obs)
## Primary splits:
## Review Rating splits as R-----------------------LL, improve=0.13383000, (0 missing)
## Previous Purchases < 3.5 to the right, improve=0.03275572, (0 missing)
## Age < 50.5 to the left, improve=0.03096766, (0 missing)
##
## Node number 109: 6 observations
## mean=85.5, MSE=137.9167
##
## Node number 130: 48 observations, complexity param=0.001136198
## mean=52.08333, MSE=505.4097
## left son=260 (22 obs) right son=261 (26 obs)
## Primary splits:
## Previous Purchases < 44.5 to the right, improve=0.064765570, (0 missing)
## Age < 51.5 to the left, improve=0.029026230, (0 missing)
## Review Rating splits as ---L-------R---------R----, improve=0.003657648, (0 missing)
## Surrogate splits:
## Age < 31.5 to the left, agree=0.583, adj=0.091, (0 split)
## Review Rating splits as ---R-------L---------R----, agree=0.562, adj=0.045, (0 split)
##
## Node number 131: 13 observations, complexity param=0.001136198
## mean=62.23077, MSE=828.7929
## left son=262 (6 obs) right son=263 (7 obs)
## Primary splits:
## Age < 49.5 to the right, improve=0.4094505, (0 missing)
## Previous Purchases < 45.5 to the left, improve=0.2576321, (0 missing)
## Surrogate splits:
## Previous Purchases < 44.5 to the right, agree=0.615, adj=0.167, (0 split)
##
## Node number 132: 73 observations, complexity param=0.001084026
## mean=52, MSE=654.8493
## left son=264 (42 obs) right son=265 (31 obs)
## Primary splits:
## Age < 48.5 to the left, improve=0.03193122, (0 missing)
## Review Rating splits as ---LR------L---------L----, improve=0.01547551, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.01301461, (0 missing)
## Surrogate splits:
## Review Rating splits as ---LL------R---------L----, agree=0.589, adj=0.032, (0 split)
##
## Node number 133: 296 observations, complexity param=0.001084026
## mean=56.31419, MSE=570.3574
## left son=266 (257 obs) right son=267 (39 obs)
## Primary splits:
## Previous Purchases < 17.5 to the right, improve=0.007843108, (0 missing)
## Age < 40.5 to the left, improve=0.004979799, (0 missing)
## Review Rating splits as ---LL------R---------L----, improve=0.001714605, (0 missing)
##
## Node number 134: 5 observations
## mean=38, MSE=91.6
##
## Node number 135: 21 observations
## mean=74.33333, MSE=497.6508
##
## Node number 160: 11 observations
## mean=29.63636, MSE=50.77686
##
## Node number 161: 15 observations
## mean=54.8, MSE=590.2933
##
## Node number 164: 14 observations, complexity param=0.001018377
## mean=49.57143, MSE=561.102
## left son=328 (8 obs) right son=329 (6 obs)
## Primary splits:
## Age < 61.5 to the right, improve=0.22341920, (0 missing)
## Review Rating splits as ------L-----R-L-----------, improve=0.06416627, (0 missing)
## Previous Purchases < 45.5 to the right, improve=0.04320943, (0 missing)
## Surrogate splits:
## Review Rating splits as ------L-----R-L-----------, agree=0.714, adj=0.333, (0 split)
##
## Node number 165: 2 observations
## mean=95, MSE=4
##
## Node number 172: 11 observations
## mean=44.27273, MSE=323.1074
##
## Node number 173: 179 observations, complexity param=0.001163895
## mean=60.1676, MSE=508.2065
## left son=346 (73 obs) right son=347 (106 obs)
## Primary splits:
## Previous Purchases < 17.5 to the left, improve=0.016957430, (0 missing)
## Age < 63.5 to the right, improve=0.011430800, (0 missing)
## Review Rating splits as --R-------L----L--L-R-R---, improve=0.008278216, (0 missing)
## Surrogate splits:
## Review Rating splits as --R-------R----R--L-R-R---, agree=0.609, adj=0.041, (0 split)
## Age < 66.5 to the right, agree=0.598, adj=0.014, (0 split)
##
## Node number 174: 17 observations
## mean=62.82353, MSE=381.5571
##
## Node number 175: 17 observations, complexity param=0.001037097
## mean=74.82353, MSE=508.4983
## left son=350 (3 obs) right son=351 (14 obs)
## Primary splits:
## Age < 60 to the left, improve=0.3832446000, (0 missing)
## Previous Purchases < 44.5 to the right, improve=0.0127295300, (0 missing)
## Review Rating splits as ----------R---------L-R---, improve=0.0006973568, (0 missing)
##
## Node number 176: 6 observations
## mean=33.33333, MSE=180.8889
##
## Node number 177: 168 observations, complexity param=0.001021136
## mean=52.93452, MSE=487.3231
## left son=354 (106 obs) right son=355 (62 obs)
## Primary splits:
## Age < 42.5 to the left, improve=0.017399170, (0 missing)
## Previous Purchases < 11.5 to the left, improve=0.015971650, (0 missing)
## Review Rating splits as ------R---L---------R-L---, improve=0.002427344, (0 missing)
## Surrogate splits:
## Previous Purchases < 1.5 to the right, agree=0.643, adj=0.032, (0 split)
##
## Node number 178: 272 observations, complexity param=0.001598063
## mean=57.02206, MSE=569.4186
## left son=356 (121 obs) right son=357 (151 obs)
## Primary splits:
## Age < 41.5 to the right, improve=0.01766243, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.01468247, (0 missing)
## Review Rating splits as -LL-----RR--R-RL--R-------, improve=0.01429853, (0 missing)
## Surrogate splits:
## Previous Purchases < 20.5 to the right, agree=0.610, adj=0.124, (0 split)
## Review Rating splits as -LR-----RR--R-RR--L-------, agree=0.585, adj=0.066, (0 split)
##
## Node number 179: 152 observations, complexity param=0.001598063
## mean=63.41447, MSE=480.4795
## left son=358 (130 obs) right son=359 (22 obs)
## Primary splits:
## Review Rating splits as -LL-----LL--L-LR--L-------, improve=0.03391488, (0 missing)
## Previous Purchases < 47.5 to the right, improve=0.03329217, (0 missing)
## Age < 36.5 to the left, improve=0.01239627, (0 missing)
##
## Node number 180: 78 observations, complexity param=0.001304644
## mean=53.0641, MSE=528.3677
## left son=360 (53 obs) right son=361 (25 obs)
## Primary splits:
## Age < 27.5 to the left, improve=0.063833300, (0 missing)
## Previous Purchases < 28.5 to the right, improve=0.047468600, (0 missing)
## Review Rating splits as ------------R--L--L-------, improve=0.004072759, (0 missing)
##
## Node number 181: 50 observations, complexity param=0.001167516
## mean=61.04, MSE=567.8384
## left son=362 (5 obs) right son=363 (45 obs)
## Primary splits:
## Previous Purchases < 48.5 to the right, improve=0.08994609, (0 missing)
## Review Rating splits as ------------L--R--R-------, improve=0.05888486, (0 missing)
## Age < 22.5 to the right, improve=0.03052901, (0 missing)
##
## Node number 182: 211 observations, complexity param=0.001326225
## mean=60.27488, MSE=498.2752
## left son=364 (145 obs) right son=365 (66 obs)
## Primary splits:
## Previous Purchases < 40.5 to the left, improve=0.028842550, (0 missing)
## Age < 30.5 to the left, improve=0.025116460, (0 missing)
## Review Rating splits as -RL---L-LRR---L-----R-R---, improve=0.007666054, (0 missing)
## Surrogate splits:
## Review Rating splits as -LL---R-LLL---L-----L-L---, agree=0.697, adj=0.03, (0 split)
##
## Node number 183: 138 observations, complexity param=0.001211694
## mean=66.03623, MSE=497.9625
## left son=366 (80 obs) right son=367 (58 obs)
## Primary splits:
## Review Rating splits as -RR---R-LLR---L-----L-L---, improve=0.021503040, (0 missing)
## Age < 20.5 to the left, improve=0.017936460, (0 missing)
## Previous Purchases < 6.5 to the left, improve=0.002469763, (0 missing)
## Surrogate splits:
## Previous Purchases < 9.5 to the left, agree=0.645, adj=0.155, (0 split)
## Age < 18.5 to the right, agree=0.587, adj=0.017, (0 split)
##
## Node number 184: 33 observations
## mean=48.57576, MSE=351.4564
##
## Node number 185: 16 observations, complexity param=0.001598063
## mean=62.3125, MSE=630.3398
## left son=370 (5 obs) right son=371 (11 obs)
## Primary splits:
## Review Rating splits as -RL------LR---RL----R-----, improve=0.70703030, (0 missing)
## Age < 51.5 to the right, improve=0.16414860, (0 missing)
## Previous Purchases < 49.5 to the left, improve=0.05695893, (0 missing)
##
## Node number 186: 104 observations
## mean=59.86538, MSE=523.328
##
## Node number 187: 47 observations, complexity param=0.001116202
## mean=67.93617, MSE=603.0385
## left son=374 (28 obs) right son=375 (19 obs)
## Primary splits:
## Previous Purchases < 33.5 to the right, improve=0.09788888, (0 missing)
## Review Rating splits as -RL------RR---RR----L-----, improve=0.09321710, (0 missing)
## Age < 49.5 to the right, improve=0.02873069, (0 missing)
## Surrogate splits:
## Review Rating splits as -RL------LL---LL----L-----, agree=0.745, adj=0.368, (0 split)
## Age < 49.5 to the right, agree=0.617, adj=0.053, (0 split)
##
## Node number 192: 25 observations
## mean=40.48, MSE=221.1296
##
## Node number 193: 5 observations
## mean=65.8, MSE=460.56
##
## Node number 196: 29 observations
## mean=50.93103, MSE=306.3401
##
## Node number 197: 349 observations, complexity param=0.001358224
## mean=60.39542, MSE=569.0127
## left son=394 (47 obs) right son=395 (302 obs)
## Primary splits:
## Previous Purchases < 42.5 to the right, improve=0.005490680, (0 missing)
## Review Rating splits as -----L-L-----L--LL-R---R--, improve=0.004280685, (0 missing)
## Age < 42.5 to the left, improve=0.003990503, (0 missing)
##
## Node number 198: 48 observations, complexity param=0.001219142
## mean=56.20833, MSE=539.7899
## left son=396 (40 obs) right son=397 (8 obs)
## Primary splits:
## Previous Purchases < 13.5 to the right, improve=0.10292070, (0 missing)
## Age < 63.5 to the right, improve=0.07677603, (0 missing)
## Review Rating splits as -------R-----------L------, improve=0.01819721, (0 missing)
##
## Node number 199: 121 observations, complexity param=0.00120641
## mean=66.71074, MSE=555.0155
## left son=398 (102 obs) right son=399 (19 obs)
## Primary splits:
## Previous Purchases < 44.5 to the left, improve=0.03887856, (0 missing)
## Age < 69.5 to the right, improve=0.02790144, (0 missing)
## Review Rating splits as -----R-------L--RL-----L--, improve=0.01191083, (0 missing)
##
## Node number 200: 37 observations, complexity param=0.001397574
## mean=51.05405, MSE=573.4025
## left son=400 (27 obs) right son=401 (10 obs)
## Primary splits:
## Age < 29.5 to the left, improve=0.26215130, (0 missing)
## Review Rating splits as -----L-R-----L--LR-R---L--, improve=0.07151912, (0 missing)
## Previous Purchases < 10.5 to the left, improve=0.02877911, (0 missing)
##
## Node number 201: 57 observations, complexity param=0.001397574
## mean=62.63158, MSE=621.3906
## left son=402 (30 obs) right son=403 (27 obs)
## Primary splits:
## Age < 24.5 to the right, improve=0.14801580, (0 missing)
## Review Rating splits as -----R-R-----R--RR-R---L--, improve=0.09734186, (0 missing)
## Previous Purchases < 7.5 to the left, improve=0.06419446, (0 missing)
## Surrogate splits:
## Previous Purchases < 3.5 to the left, agree=0.596, adj=0.148, (0 split)
## Review Rating splits as -----L-L-----R--RL-R---L--, agree=0.579, adj=0.111, (0 split)
##
## Node number 214: 73 observations
## mean=66.63014, MSE=422.8906
##
## Node number 215: 4 observations
## mean=92, MSE=37.5
##
## Node number 216: 32 observations, complexity param=0.001053613
## mean=58.5, MSE=550.8125
## left son=432 (24 obs) right son=433 (8 obs)
## Primary splits:
## Age < 54.5 to the left, improve=0.060516660, (0 missing)
## Previous Purchases < 4.5 to the right, improve=0.043831350, (0 missing)
## Review Rating splits as ------------------------RL, improve=0.006940161, (0 missing)
## Surrogate splits:
## Previous Purchases < 6.5 to the left, agree=0.812, adj=0.25, (0 split)
##
## Node number 217: 7 observations
## mean=80.85714, MSE=136.4082
##
## Node number 260: 22 observations
## mean=45.86364, MSE=409.936
##
## Node number 261: 26 observations, complexity param=0.001136198
## mean=57.34615, MSE=525.7648
## left son=522 (18 obs) right son=523 (8 obs)
## Primary splits:
## Age < 51.5 to the left, improve=0.32238820, (0 missing)
## Previous Purchases < 42.5 to the left, improve=0.05617345, (0 missing)
## Review Rating splits as ---R-------L---------R----, improve=0.02573719, (0 missing)
##
## Node number 262: 6 observations, complexity param=0.001136198
## mean=42.33333, MSE=686.5556
## left son=524 (5 obs) right son=525 (1 obs)
## Primary splits:
## Previous Purchases < 45.5 to the left, improve=0.9354264, (0 missing)
## Age < 65 to the left, improve=0.2816799, (0 missing)
##
## Node number 263: 7 observations
## mean=79.28571, MSE=320.4898
##
## Node number 264: 42 observations, complexity param=0.001069367
## mean=48.07143, MSE=621.0663
## left son=528 (7 obs) right son=529 (35 obs)
## Primary splits:
## Review Rating splits as ---RR------L---------R----, improve=0.05623146, (0 missing)
## Age < 40.5 to the right, improve=0.04437236, (0 missing)
## Previous Purchases < 11.5 to the right, improve=0.04191964, (0 missing)
##
## Node number 265: 31 observations, complexity param=0.001084026
## mean=57.32258, MSE=651.3798
## left son=530 (10 obs) right son=531 (21 obs)
## Primary splits:
## Review Rating splits as ---LR------R---------L----, improve=0.16498160, (0 missing)
## Age < 65.5 to the right, improve=0.05522074, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.04950273, (0 missing)
##
## Node number 266: 257 observations, complexity param=0.001084026
## mean=55.49027, MSE=571.5495
## left son=532 (113 obs) right son=533 (144 obs)
## Primary splits:
## Age < 40.5 to the left, improve=0.006741811, (0 missing)
## Previous Purchases < 33.5 to the right, improve=0.004072488, (0 missing)
## Review Rating splits as ---LL------R---------L----, improve=0.001888334, (0 missing)
## Surrogate splits:
## Previous Purchases < 20.5 to the left, agree=0.584, adj=0.053, (0 split)
##
## Node number 267: 39 observations, complexity param=0.001084026
## mean=61.74359, MSE=528.5496
## left son=534 (1 obs) right son=535 (38 obs)
## Primary splits:
## Age < 20.5 to the left, improve=0.07864386, (0 missing)
## Previous Purchases < 16.5 to the left, improve=0.02472995, (0 missing)
## Review Rating splits as ---LR------R---------R----, improve=0.01266838, (0 missing)
##
## Node number 328: 8 observations
## mean=39.875, MSE=195.3594
##
## Node number 329: 6 observations, complexity param=0.001018377
## mean=62.5, MSE=756.25
## left son=658 (4 obs) right son=659 (2 obs)
## Primary splits:
## Age < 59.5 to the left, improve=0.595041300, (0 missing)
## Previous Purchases < 48.5 to the left, improve=0.136675800, (0 missing)
## Review Rating splits as ------L-----R-------------, improve=0.001487603, (0 missing)
## Surrogate splits:
## Previous Purchases < 48.5 to the left, agree=0.833, adj=0.5, (0 split)
##
## Node number 346: 73 observations, complexity param=0.001163895
## mean=56.63014, MSE=421.192
## left son=692 (43 obs) right son=693 (30 obs)
## Primary splits:
## Review Rating splits as --R-------L----L--L-R-R---, improve=0.110561100, (0 missing)
## Previous Purchases < 14.5 to the right, improve=0.043450520, (0 missing)
## Age < 63.5 to the right, improve=0.008788439, (0 missing)
## Surrogate splits:
## Age < 58.5 to the right, agree=0.603, adj=0.033, (0 split)
## Previous Purchases < 3.5 to the right, agree=0.603, adj=0.033, (0 split)
##
## Node number 347: 106 observations
## mean=62.60377, MSE=553.5789
##
## Node number 350: 3 observations
## mean=44.66667, MSE=96.22222
##
## Node number 351: 14 observations
## mean=81.28571, MSE=360.2041
##
## Node number 354: 106 observations, complexity param=0.001021136
## mean=50.70755, MSE=467.1315
## left son=708 (24 obs) right son=709 (82 obs)
## Primary splits:
## Previous Purchases < 11.5 to the left, improve=0.061605690, (0 missing)
## Age < 37.5 to the right, improve=0.007096820, (0 missing)
## Review Rating splits as ------L---L---------R-L---, improve=0.003279954, (0 missing)
##
## Node number 355: 62 observations
## mean=56.74194, MSE=498.8689
##
## Node number 356: 121 observations, complexity param=0.001080582
## mean=53.47934, MSE=544.9768
## left son=712 (113 obs) right son=713 (8 obs)
## Primary splits:
## Previous Purchases < 2.5 to the right, improve=0.030791450, (0 missing)
## Review Rating splits as -LL-----RL--L-RR--R-------, improve=0.018550380, (0 missing)
## Age < 42.5 to the left, improve=0.009241003, (0 missing)
##
## Node number 357: 151 observations, complexity param=0.001598063
## mean=59.86093, MSE=570.8879
## left son=714 (44 obs) right son=715 (107 obs)
## Primary splits:
## Age < 34.5 to the left, improve=0.05284704, (0 missing)
## Review Rating splits as -RL-----RR--R-LL--R-------, improve=0.03874871, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.01047699, (0 missing)
## Surrogate splits:
## Review Rating splits as -RL-----RR--R-RR--R-------, agree=0.715, adj=0.023, (0 split)
##
## Node number 358: 130 observations, complexity param=0.001598063
## mean=61.75385, MSE=451.2163
## left son=716 (21 obs) right son=717 (109 obs)
## Primary splits:
## Previous Purchases < 47.5 to the right, improve=0.083591850, (0 missing)
## Age < 41.5 to the right, improve=0.015795180, (0 missing)
## Review Rating splits as -LR-----LL--R-L---R-------, improve=0.009795024, (0 missing)
##
## Node number 359: 22 observations, complexity param=0.001082269
## mean=73.22727, MSE=540.812
## left son=718 (16 obs) right son=719 (6 obs)
## Primary splits:
## Previous Purchases < 45 to the left, improve=0.1989668, (0 missing)
## Age < 35.5 to the left, improve=0.1572787, (0 missing)
##
## Node number 360: 53 observations, complexity param=0.001304644
## mean=49.07547, MSE=450.749
## left son=720 (15 obs) right son=721 (38 obs)
## Primary splits:
## Previous Purchases < 6.5 to the left, improve=0.101054200, (0 missing)
## Age < 22.5 to the left, improve=0.044456620, (0 missing)
## Review Rating splits as ------------L--R--L-------, improve=0.002184139, (0 missing)
##
## Node number 361: 25 observations, complexity param=0.001051063
## mean=61.52, MSE=587.6896
## left son=722 (2 obs) right son=723 (23 obs)
## Primary splits:
## Previous Purchases < 28.5 to the right, improve=0.15647870, (0 missing)
## Age < 30.5 to the right, improve=0.04642453, (0 missing)
## Review Rating splits as ------------R--L--L-------, improve=0.01454340, (0 missing)
##
## Node number 362: 5 observations, complexity param=0.001027371
## mean=39.6, MSE=464.24
## left son=724 (4 obs) right son=725 (1 obs)
## Primary splits:
## Age < 21 to the right, improve=0.9681199, (0 missing)
## Review Rating splits as ------------R--L--L-------, improve=0.4601212, (0 missing)
## Previous Purchases < 49.5 to the left, improve=0.1062094, (0 missing)
##
## Node number 363: 45 observations, complexity param=0.00115575
## mean=63.42222, MSE=522.5995
## left son=726 (10 obs) right son=727 (35 obs)
## Primary splits:
## Review Rating splits as ------------L--R--R-------, improve=0.10749710, (0 missing)
## Previous Purchases < 47.5 to the left, improve=0.07714130, (0 missing)
## Age < 30.5 to the right, improve=0.01265048, (0 missing)
##
## Node number 364: 145 observations
## mean=57.71724, MSE=489.4718
##
## Node number 365: 66 observations, complexity param=0.001326225
## mean=65.89394, MSE=471.6706
## left son=730 (29 obs) right son=731 (37 obs)
## Primary splits:
## Review Rating splits as -RL---L-LRR---L-----R-R---, improve=0.09556695, (0 missing)
## Age < 30.5 to the left, improve=0.03829311, (0 missing)
## Previous Purchases < 44.5 to the right, improve=0.02693985, (0 missing)
## Surrogate splits:
## Age < 25.5 to the right, agree=0.636, adj=0.172, (0 split)
## Previous Purchases < 45.5 to the right, agree=0.576, adj=0.034, (0 split)
##
## Node number 366: 80 observations, complexity param=0.001211694
## mean=63.25, MSE=506.1875
## left son=732 (33 obs) right son=733 (47 obs)
## Primary splits:
## Age < 24.5 to the left, improve=0.094408820, (0 missing)
## Previous Purchases < 11.5 to the right, improve=0.055772630, (0 missing)
## Review Rating splits as --------RR----R-----L-R---, improve=0.004657697, (0 missing)
##
## Node number 367: 58 observations
## mean=69.87931, MSE=461.1406
##
## Node number 370: 5 observations
## mean=31, MSE=82.8
##
## Node number 371: 11 observations
## mean=76.54545, MSE=230.9752
##
## Node number 374: 28 observations, complexity param=0.001097233
## mean=61.60714, MSE=611.2385
## left son=748 (7 obs) right son=749 (21 obs)
## Primary splits:
## Review Rating splits as -RR------RR---RR----L-----, improve=0.14023120, (0 missing)
## Previous Purchases < 34.5 to the left, improve=0.10331790, (0 missing)
## Age < 55.5 to the left, improve=0.03841384, (0 missing)
##
## Node number 375: 19 observations
## mean=77.26316, MSE=444.9307
##
## Node number 394: 47 observations, complexity param=0.001358224
## mean=55.91489, MSE=639.0991
## left son=788 (29 obs) right son=789 (18 obs)
## Primary splits:
## Review Rating splits as -----L-R-----R--LR-L---L--, improve=0.18814270, (0 missing)
## Previous Purchases < 47.5 to the left, improve=0.06272018, (0 missing)
## Age < 49.5 to the left, improve=0.05175054, (0 missing)
## Surrogate splits:
## Age < 52.5 to the left, agree=0.681, adj=0.167, (0 split)
##
## Node number 395: 302 observations, complexity param=0.00129495
## mean=61.09272, MSE=554.4947
## left son=790 (179 obs) right son=791 (123 obs)
## Primary splits:
## Review Rating splits as -----R-L-----L--LL-R---R--, improve=0.016410310, (0 missing)
## Previous Purchases < 32.5 to the left, improve=0.005779130, (0 missing)
## Age < 51.5 to the right, improve=0.003512665, (0 missing)
## Surrogate splits:
## Previous Purchases < 41.5 to the left, agree=0.599, adj=0.016, (0 split)
##
## Node number 396: 40 observations, complexity param=0.001007815
## mean=52.875, MSE=455.0594
## left son=792 (27 obs) right son=793 (13 obs)
## Primary splits:
## Age < 63.5 to the right, improve=0.09880456, (0 missing)
## Review Rating splits as -------R-----------L------, improve=0.05414687, (0 missing)
## Previous Purchases < 33.5 to the right, improve=0.05383922, (0 missing)
## Surrogate splits:
## Previous Purchases < 47 to the left, agree=0.725, adj=0.154, (0 split)
##
## Node number 397: 8 observations
## mean=72.875, MSE=630.1094
##
## Node number 398: 102 observations, complexity param=0.00120641
## mean=64.70588, MSE=580.8939
## left son=796 (3 obs) right son=797 (99 obs)
## Primary splits:
## Previous Purchases < 43.5 to the right, improve=0.04500622, (0 missing)
## Age < 69.5 to the right, improve=0.03388359, (0 missing)
## Review Rating splits as -----R-------L--RL-----L--, improve=0.02156070, (0 missing)
##
## Node number 399: 19 observations
## mean=77.47368, MSE=278.6704
##
## Node number 400: 27 observations
## mean=43.59259, MSE=431.9451
##
## Node number 401: 10 observations
## mean=71.2, MSE=399.16
##
## Node number 402: 30 observations
## mean=53.53333, MSE=448.8489
##
## Node number 403: 27 observations, complexity param=0.001397574
## mean=72.74074, MSE=618.9328
## left son=806 (14 obs) right son=807 (13 obs)
## Primary splits:
## Review Rating splits as -----R-R-----L--RR-L---L--, improve=0.20881810, (0 missing)
## Previous Purchases < 7.5 to the left, improve=0.11041750, (0 missing)
## Age < 18.5 to the right, improve=0.04655041, (0 missing)
## Surrogate splits:
## Age < 20.5 to the left, agree=0.704, adj=0.385, (0 split)
## Previous Purchases < 6.5 to the left, agree=0.630, adj=0.231, (0 split)
##
## Node number 432: 24 observations
## mean=55.16667, MSE=488.9722
##
## Node number 433: 8 observations, complexity param=0.001053613
## mean=68.5, MSE=603
## left son=866 (3 obs) right son=867 (5 obs)
## Primary splits:
## Previous Purchases < 5 to the right, improve=0.73435600, (0 missing)
## Age < 55.5 to the right, improve=0.22038620, (0 missing)
## Review Rating splits as ------------------------RL, improve=0.03701729, (0 missing)
## Surrogate splits:
## Age < 60.5 to the right, agree=0.75, adj=0.333, (0 split)
## Review Rating splits as ------------------------RL, agree=0.75, adj=0.333, (0 split)
##
## Node number 522: 18 observations
## mean=48.66667, MSE=435.6667
##
## Node number 523: 8 observations
## mean=76.875, MSE=177.6094
##
## Node number 524: 5 observations
## mean=31, MSE=53.2
##
## Node number 525: 1 observations
## mean=99, MSE=0
##
## Node number 528: 7 observations
## mean=34.85714, MSE=182.1224
##
## Node number 529: 35 observations, complexity param=0.001069367
## mean=50.71429, MSE=666.9469
## left son=1058 (7 obs) right son=1059 (28 obs)
## Primary splits:
## Age < 44.5 to the right, improve=0.055270130, (0 missing)
## Previous Purchases < 11.5 to the right, improve=0.021955480, (0 missing)
## Review Rating splits as ---RL----------------R----, improve=0.007108848, (0 missing)
##
## Node number 530: 10 observations
## mean=42.3, MSE=213.01
##
## Node number 531: 21 observations, complexity param=0.001038539
## mean=64.47619, MSE=701.4875
## left son=1062 (3 obs) right son=1063 (18 obs)
## Primary splits:
## Age < 65.5 to the right, improve=0.1542048000, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.0494191200, (0 missing)
## Review Rating splits as ----R------L--------------, improve=0.0008359799, (0 missing)
##
## Node number 532: 113 observations
## mean=53.27434, MSE=543.792
##
## Node number 533: 144 observations, complexity param=0.001084026
## mean=57.22917, MSE=586.4544
## left son=1066 (125 obs) right son=1067 (19 obs)
## Primary splits:
## Age < 44.5 to the right, improve=0.070628690, (0 missing)
## Review Rating splits as ---LL------R---------L----, improve=0.007408687, (0 missing)
## Previous Purchases < 33.5 to the right, improve=0.002946523, (0 missing)
##
## Node number 534: 1 observations
## mean=22, MSE=0
##
## Node number 535: 38 observations, complexity param=0.001084026
## mean=62.78947, MSE=499.7978
## left son=1070 (35 obs) right son=1071 (3 obs)
## Primary splits:
## Age < 27.5 to the right, improve=0.17426860, (0 missing)
## Review Rating splits as ---LR------R---------R----, improve=0.02367406, (0 missing)
## Previous Purchases < 16.5 to the left, improve=0.01769314, (0 missing)
##
## Node number 658: 4 observations
## mean=47.5, MSE=456.25
##
## Node number 659: 2 observations
## mean=92.5, MSE=6.25
##
## Node number 692: 43 observations
## mean=50.93023, MSE=313.9254
##
## Node number 693: 30 observations, complexity param=0.001163895
## mean=64.8, MSE=461.6267
## left son=1386 (5 obs) right son=1387 (25 obs)
## Primary splits:
## Previous Purchases < 13.5 to the right, improve=0.26218590, (0 missing)
## Age < 65.5 to the right, improve=0.07120713, (0 missing)
## Review Rating splits as --L-----------------R-L---, improve=0.01482800, (0 missing)
##
## Node number 708: 24 observations, complexity param=0.001021136
## mean=40.79167, MSE=414.1649
## left son=1416 (17 obs) right son=1417 (7 obs)
## Primary splits:
## Previous Purchases < 5.5 to the right, improve=0.2299530, (0 missing)
## Age < 40 to the left, improve=0.1186122, (0 missing)
## Review Rating splits as ------R---R---------L-L---, improve=0.1177070, (0 missing)
##
## Node number 709: 82 observations
## mean=53.60976, MSE=445.4331
##
## Node number 712: 113 observations, complexity param=0.001080582
## mean=52.38938, MSE=528.344
## left son=1424 (45 obs) right son=1425 (68 obs)
## Primary splits:
## Review Rating splits as -LL-----RL--R-RR--R-------, improve=0.030079190, (0 missing)
## Previous Purchases < 20.5 to the left, improve=0.009545057, (0 missing)
## Age < 42.5 to the left, improve=0.006563998, (0 missing)
## Surrogate splits:
## Age < 42.5 to the left, agree=0.619, adj=0.044, (0 split)
##
## Node number 713: 8 observations, complexity param=0.001080582
## mean=68.875, MSE=526.1094
## left son=1426 (2 obs) right son=1427 (6 obs)
## Primary splits:
## Review Rating splits as -R----------L-LR--R-------, improve=0.6236920, (0 missing)
## Previous Purchases < 1.5 to the left, improve=0.4207795, (0 missing)
## Age < 47.5 to the left, improve=0.2630538, (0 missing)
## Surrogate splits:
## Previous Purchases < 1.5 to the left, agree=0.875, adj=0.5, (0 split)
##
## Node number 714: 44 observations, complexity param=0.001387775
## mean=51.29545, MSE=632.5263
## left son=1428 (11 obs) right son=1429 (33 obs)
## Primary splits:
## Review Rating splits as -RR-----RR--R-LL--R-------, improve=0.10906920, (0 missing)
## Age < 32.5 to the right, improve=0.06846393, (0 missing)
## Previous Purchases < 13.5 to the right, improve=0.05049510, (0 missing)
##
## Node number 715: 107 observations, complexity param=0.001293294
## mean=63.38318, MSE=502.9653
## left son=1430 (26 obs) right son=1431 (81 obs)
## Primary splits:
## Review Rating splits as -RL-----RR--R-RL--R-------, improve=0.05256419, (0 missing)
## Age < 37.5 to the right, improve=0.02660337, (0 missing)
## Previous Purchases < 2 to the left, improve=0.01398953, (0 missing)
##
## Node number 716: 21 observations
## mean=47.7619, MSE=231.8005
##
## Node number 717: 109 observations
## mean=64.44954, MSE=448.5043
##
## Node number 718: 16 observations
## mean=66.875, MSE=593.9844
##
## Node number 719: 6 observations
## mean=90.16667, MSE=4.472222
##
## Node number 720: 15 observations
## mean=38.33333, MSE=256.6222
##
## Node number 721: 38 observations, complexity param=0.001304644
## mean=53.31579, MSE=463.8476
## left son=1442 (15 obs) right son=1443 (23 obs)
## Primary splits:
## Age < 22.5 to the left, improve=0.12027880, (0 missing)
## Previous Purchases < 8.5 to the right, improve=0.05849365, (0 missing)
## Review Rating splits as ------------R--R--L-------, improve=0.02358318, (0 missing)
## Surrogate splits:
## Previous Purchases < 29.5 to the right, agree=0.632, adj=0.067, (0 split)
##
## Node number 722: 2 observations
## mean=29, MSE=64
##
## Node number 723: 23 observations
## mean=64.34783, MSE=533.2703
##
## Node number 724: 4 observations
## mean=29, MSE=18.5
##
## Node number 725: 1 observations
## mean=82, MSE=0
##
## Node number 726: 10 observations
## mean=49.4, MSE=301.64
##
## Node number 727: 35 observations
## mean=67.42857, MSE=513.502
##
## Node number 730: 29 observations
## mean=58.31034, MSE=506.7658
##
## Node number 731: 37 observations
## mean=71.83784, MSE=363.7575
##
## Node number 732: 33 observations
## mean=55, MSE=457.7576
##
## Node number 733: 47 observations, complexity param=0.001211694
## mean=69.04255, MSE=458.8493
## left son=1466 (5 obs) right son=1467 (42 obs)
## Primary splits:
## Previous Purchases < 15.5 to the right, improve=0.11055480, (0 missing)
## Age < 26.5 to the right, improve=0.08818112, (0 missing)
## Review Rating splits as --------LR----L-----L-L---, improve=0.03740713, (0 missing)
##
## Node number 748: 7 observations
## mean=45.57143, MSE=547.6735
##
## Node number 749: 21 observations
## mean=66.95238, MSE=518.1406
##
## Node number 788: 29 observations
## mean=47.27586, MSE=538.6825
##
## Node number 789: 18 observations, complexity param=0.001007866
## mean=69.83333, MSE=486.9167
## left son=1578 (4 obs) right son=1579 (14 obs)
## Primary splits:
## Age < 56 to the right, improve=2.250352e-01, (0 missing)
## Previous Purchases < 47.5 to the left, improve=1.208029e-01, (0 missing)
## Review Rating splits as -------R-----L---L--------, improve=4.563866e-05, (0 missing)
##
## Node number 790: 179 observations, complexity param=0.00103108
## mean=58.59218, MSE=549.5264
## left son=1580 (10 obs) right son=1581 (169 obs)
## Primary splits:
## Previous Purchases < 9.5 to the left, improve=0.022927940, (0 missing)
## Age < 50.5 to the right, improve=0.012443230, (0 missing)
## Review Rating splits as -------L-----L--RL--------, improve=0.000585263, (0 missing)
##
## Node number 791: 123 observations, complexity param=0.00129495
## mean=64.73171, MSE=539.3833
## left son=1582 (101 obs) right son=1583 (22 obs)
## Primary splits:
## Previous Purchases < 13.5 to the right, improve=0.039254310, (0 missing)
## Age < 45.5 to the right, improve=0.024171570, (0 missing)
## Review Rating splits as -----L-------------R---R--, improve=0.007001794, (0 missing)
##
## Node number 792: 27 observations, complexity param=0.001007815
## mean=48.22222, MSE=382.0988
## left son=1584 (11 obs) right son=1585 (16 obs)
## Primary splits:
## Review Rating splits as -------R-----------L------, improve=0.25302470, (0 missing)
## Previous Purchases < 19 to the left, improve=0.08570485, (0 missing)
## Age < 66.5 to the left, improve=0.03882999, (0 missing)
## Surrogate splits:
## Age < 65.5 to the left, agree=0.704, adj=0.273, (0 split)
## Previous Purchases < 29.5 to the right, agree=0.667, adj=0.182, (0 split)
##
## Node number 793: 13 observations
## mean=62.53846, MSE=468.2485
##
## Node number 796: 3 observations
## mean=35.33333, MSE=24.22222
##
## Node number 797: 99 observations, complexity param=0.001039042
## mean=65.59596, MSE=570.8267
## left son=1594 (9 obs) right son=1595 (90 obs)
## Primary splits:
## Age < 69.5 to the right, improve=0.04021683, (0 missing)
## Previous Purchases < 14.5 to the right, improve=0.03888308, (0 missing)
## Review Rating splits as -----R-------L--LL-----L--, improve=0.01596003, (0 missing)
##
## Node number 806: 14 observations, complexity param=0.001397574
## mean=61.78571, MSE=789.3112
## left son=1612 (12 obs) right son=1613 (2 obs)
## Primary splits:
## Previous Purchases < 2.5 to the right, improve=0.300339400, (0 missing)
## Age < 18.5 to the right, improve=0.182257200, (0 missing)
## Review Rating splits as -------------R-----L---L--, improve=0.007271905, (0 missing)
##
## Node number 807: 13 observations
## mean=84.53846, MSE=167.0178
##
## Node number 866: 3 observations
## mean=41.33333, MSE=140.2222
##
## Node number 867: 5 observations
## mean=84.8, MSE=172.16
##
## Node number 1058: 7 observations, complexity param=0.001069367
## mean=38.57143, MSE=663.6735
## left son=2116 (6 obs) right son=2117 (1 obs)
## Primary splits:
## Review Rating splits as ---LL----------------R----, improve=0.9170203, (0 missing)
## Previous Purchases < 13.5 to the left, improve=0.1308836, (0 missing)
## Age < 46.5 to the left, improve=0.1193389, (0 missing)
##
## Node number 1059: 28 observations
## mean=53.75, MSE=621.6875
##
## Node number 1062: 3 observations
## mean=39, MSE=28.66667
##
## Node number 1063: 18 observations
## mean=68.72222, MSE=687.4228
##
## Node number 1066: 125 observations
## mean=54.72, MSE=565.2096
##
## Node number 1067: 19 observations
## mean=73.73684, MSE=412.2992
##
## Node number 1070: 35 observations
## mean=60.05714, MSE=446.6253
##
## Node number 1071: 3 observations
## mean=94.66667, MSE=16.88889
##
## Node number 1386: 5 observations
## mean=40.2, MSE=291.36
##
## Node number 1387: 25 observations
## mean=69.72, MSE=350.4416
##
## Node number 1416: 17 observations
## mean=34.52941, MSE=188.8374
##
## Node number 1417: 7 observations, complexity param=0.001021136
## mean=56, MSE=634.8571
## left son=2834 (3 obs) right son=2835 (4 obs)
## Primary splits:
## Previous Purchases < 3.5 to the left, improve=0.69950120, (0 missing)
## Age < 39.5 to the left, improve=0.24270550, (0 missing)
## Review Rating splits as ------R---L---------------, improve=0.03360336, (0 missing)
## Surrogate splits:
## Age < 39.5 to the left, agree=0.857, adj=0.667, (0 split)
## Review Rating splits as ------R---L---------------, agree=0.714, adj=0.333, (0 split)
##
## Node number 1424: 45 observations
## mean=47.48889, MSE=444.161
##
## Node number 1425: 68 observations, complexity param=0.001080582
## mean=55.63235, MSE=557.6442
## left son=2850 (8 obs) right son=2851 (60 obs)
## Primary splits:
## Age < 42.5 to the left, improve=0.057498540, (0 missing)
## Previous Purchases < 20 to the left, improve=0.044398730, (0 missing)
## Review Rating splits as --------R---L-RL--L-------, improve=0.008279087, (0 missing)
##
## Node number 1426: 2 observations
## mean=37.5, MSE=72.25
##
## Node number 1427: 6 observations
## mean=79.33333, MSE=239.8889
##
## Node number 1428: 11 observations
## mean=36.90909, MSE=125.719
##
## Node number 1429: 33 observations, complexity param=0.001340273
## mean=56.09091, MSE=709.4766
## left son=2858 (12 obs) right son=2859 (21 obs)
## Primary splits:
## Previous Purchases < 17 to the right, improve=0.10665750, (0 missing)
## Review Rating splits as -LL-----LL--R-----L-------, improve=0.08408860, (0 missing)
## Age < 33.5 to the right, improve=0.02995032, (0 missing)
## Surrogate splits:
## Review Rating splits as -LR-----RR--R-----R-------, agree=0.667, adj=0.083, (0 split)
##
## Node number 1430: 26 observations
## mean=54.30769, MSE=488.213
##
## Node number 1431: 81 observations
## mean=66.2963, MSE=472.7764
##
## Node number 1442: 15 observations
## mean=44.06667, MSE=227.2622
##
## Node number 1443: 23 observations, complexity param=0.001304644
## mean=59.34783, MSE=525.966
## left son=2886 (12 obs) right son=2887 (11 obs)
## Primary splits:
## Age < 25.5 to the right, improve=0.42697570, (0 missing)
## Previous Purchases < 28.5 to the right, improve=0.11046760, (0 missing)
## Review Rating splits as ------------R--L--L-------, improve=0.01912508, (0 missing)
## Surrogate splits:
## Review Rating splits as ------------R--L--L-------, agree=0.652, adj=0.273, (0 split)
## Previous Purchases < 25 to the right, agree=0.652, adj=0.273, (0 split)
##
## Node number 1466: 5 observations, complexity param=0.001211694
## mean=48.4, MSE=906.64
## left son=2932 (3 obs) right son=2933 (2 obs)
## Primary splits:
## Age < 29.5 to the right, improve=0.9059164, (0 missing)
## Review Rating splits as --------R-----L-----R-L---, improve=0.3856069, (0 missing)
## Previous Purchases < 16.5 to the left, improve=0.1262794, (0 missing)
##
## Node number 1467: 42 observations
## mean=71.5, MSE=348.7738
##
## Node number 1578: 4 observations, complexity param=0.001007866
## mean=50.25, MSE=640.6875
## left son=3156 (3 obs) right son=3157 (1 obs)
## Primary splits:
## Previous Purchases < 47.5 to the left, improve=0.95083410, (0 missing)
## Review Rating splits as -------L-----L---R--------, improve=0.46444250, (0 missing)
## Age < 57.5 to the left, improve=0.03541118, (0 missing)
##
## Node number 1579: 14 observations
## mean=75.42857, MSE=302.102
##
## Node number 1580: 10 observations
## mean=44, MSE=365.8
##
## Node number 1581: 169 observations
## mean=59.45562, MSE=547.0528
##
## Node number 1582: 101 observations, complexity param=0.00129495
## mean=62.58416, MSE=548.0647
## left son=3164 (51 obs) right son=3165 (50 obs)
## Primary splits:
## Previous Purchases < 29.5 to the left, improve=0.056817950, (0 missing)
## Age < 45.5 to the right, improve=0.030105010, (0 missing)
## Review Rating splits as -----L-------------R---R--, improve=0.004803122, (0 missing)
## Surrogate splits:
## Age < 50.5 to the left, agree=0.594, adj=0.18, (0 split)
## Review Rating splits as -----R-------------L---L--, agree=0.515, adj=0.02, (0 split)
##
## Node number 1583: 22 observations
## mean=74.59091, MSE=381.1508
##
## Node number 1584: 11 observations
## mean=36.36364, MSE=172.2314
##
## Node number 1585: 16 observations, complexity param=0.001007815
## mean=56.375, MSE=363.2344
## left son=3170 (4 obs) right son=3171 (12 obs)
## Primary splits:
## Previous Purchases < 19 to the left, improve=0.4391248, (0 missing)
## Age < 67.5 to the right, improve=0.2482717, (0 missing)
## Surrogate splits:
## Age < 69.5 to the right, agree=0.812, adj=0.25, (0 split)
##
## Node number 1594: 9 observations, complexity param=0.001014141
## mean=50.44444, MSE=566.0247
## left son=3188 (7 obs) right son=3189 (2 obs)
## Primary splits:
## Previous Purchases < 15 to the right, improve=0.3426867, (0 missing)
## Review Rating splits as -----R-------L--LR-----L--, improve=0.2892438, (0 missing)
##
## Node number 1595: 90 observations
## mean=67.11111, MSE=546.0543
##
## Node number 1612: 12 observations
## mean=55.5, MSE=644.25
##
## Node number 1613: 2 observations
## mean=99.5, MSE=0.25
##
## Node number 2116: 6 observations
## mean=28.5, MSE=64.25
##
## Node number 2117: 1 observations
## mean=99, MSE=0
##
## Node number 2834: 3 observations
## mean=31.66667, MSE=46.22222
##
## Node number 2835: 4 observations
## mean=74.25, MSE=299.1875
##
## Node number 2850: 8 observations
## mean=40.125, MSE=231.1094
##
## Node number 2851: 60 observations, complexity param=0.001080582
## mean=57.7, MSE=564.8433
## left son=5702 (27 obs) right son=5703 (33 obs)
## Primary splits:
## Previous Purchases < 20 to the left, improve=0.05468737, (0 missing)
## Age < 43.5 to the right, improve=0.04550087, (0 missing)
## Review Rating splits as --------R---L-RL--R-------, improve=0.02270647, (0 missing)
## Surrogate splits:
## Age < 47.5 to the left, agree=0.567, adj=0.037, (0 split)
## Review Rating splits as --------R---R-RR--L-------, agree=0.567, adj=0.037, (0 split)
##
## Node number 2858: 12 observations
## mean=44.58333, MSE=499.2431
##
## Node number 2859: 21 observations, complexity param=0.001340273
## mean=62.66667, MSE=710.6984
## left son=5718 (14 obs) right son=5719 (7 obs)
## Primary splits:
## Age < 32.5 to the right, improve=0.2255391, (0 missing)
## Review Rating splits as -RR-----LL--R-----R-------, improve=0.1991781, (0 missing)
## Previous Purchases < 8 to the left, improve=0.1200697, (0 missing)
## Surrogate splits:
## Review Rating splits as -LL-----LL--L-----R-------, agree=0.810, adj=0.429, (0 split)
## Previous Purchases < 14.5 to the left, agree=0.762, adj=0.286, (0 split)
##
## Node number 2886: 12 observations
## mean=45, MSE=269.6667
##
## Node number 2887: 11 observations
## mean=75, MSE=336
##
## Node number 2932: 3 observations
## mean=25, MSE=2
##
## Node number 2933: 2 observations
## mean=83.5, MSE=210.25
##
## Node number 3156: 3 observations
## mean=36, MSE=42
##
## Node number 3157: 1 observations
## mean=93, MSE=0
##
## Node number 3164: 51 observations, complexity param=0.001091274
## mean=57.05882, MSE=493.3103
## left son=6328 (35 obs) right son=6329 (16 obs)
## Primary splits:
## Age < 45 to the right, improve=0.07512276, (0 missing)
## Previous Purchases < 25.5 to the right, improve=0.03401797, (0 missing)
## Review Rating splits as -----L-------------R---L--, improve=0.02039367, (0 missing)
## Surrogate splits:
## Previous Purchases < 18.5 to the right, agree=0.706, adj=0.063, (0 split)
##
## Node number 3165: 50 observations
## mean=68.22, MSE=541.0116
##
## Node number 3170: 4 observations
## mean=34.5, MSE=29.25
##
## Node number 3171: 12 observations
## mean=63.66667, MSE=261.8889
##
## Node number 3188: 7 observations, complexity param=0.001014141
## mean=43, MSE=478.2857
## left son=6376 (5 obs) right son=6377 (2 obs)
## Primary splits:
## Review Rating splits as -----L-------L--LR-----L--, improve=0.8037037, (0 missing)
## Previous Purchases < 42.5 to the left, improve=0.7056452, (0 missing)
##
## Node number 3189: 2 observations
## mean=76.5, MSE=0.25
##
## Node number 5702: 27 observations
## mean=51.55556, MSE=477.358
##
## Node number 5703: 33 observations, complexity param=0.001080582
## mean=62.72727, MSE=580.259
## left son=11406 (15 obs) right son=11407 (18 obs)
## Primary splits:
## Age < 46.5 to the right, improve=0.11789930, (0 missing)
## Previous Purchases < 27.5 to the right, improve=0.07358063, (0 missing)
## Review Rating splits as --------L---L-RL--L-------, improve=0.05678919, (0 missing)
## Surrogate splits:
## Review Rating splits as --------R---R-RL--R-------, agree=0.576, adj=0.067, (0 split)
##
## Node number 5718: 14 observations, complexity param=0.001198202
## mean=53.71429, MSE=627.0612
## left son=11436 (7 obs) right son=11437 (7 obs)
## Primary splits:
## Previous Purchases < 8.5 to the left, improve=0.27546700, (0 missing)
## Review Rating splits as -RR-----LL--R-------------, improve=0.26951120, (0 missing)
## Age < 33.5 to the right, improve=0.05127647, (0 missing)
## Surrogate splits:
## Review Rating splits as -RR-----RL--R-------------, agree=0.714, adj=0.429, (0 split)
## Age < 33.5 to the left, agree=0.643, adj=0.286, (0 split)
##
## Node number 5719: 7 observations
## mean=80.57143, MSE=397.102
##
## Node number 6328: 35 observations
## mean=52.94286, MSE=436.2824
##
## Node number 6329: 16 observations, complexity param=0.001091274
## mean=66.0625, MSE=499.9336
## left son=12658 (9 obs) right son=12659 (7 obs)
## Primary splits:
## Age < 41.5 to the left, improve=0.36054210, (0 missing)
## Previous Purchases < 16.5 to the right, improve=0.09400788, (0 missing)
## Review Rating splits as -----R-------------L---L--, improve=0.04895096, (0 missing)
## Surrogate splits:
## Review Rating splits as -----R-------------L---L--, agree=0.750, adj=0.429, (0 split)
## Previous Purchases < 16.5 to the right, agree=0.625, adj=0.143, (0 split)
##
## Node number 6376: 5 observations
## mean=30.6, MSE=53.04
##
## Node number 6377: 2 observations
## mean=74, MSE=196
##
## Node number 11406: 15 observations, complexity param=0.001080582
## mean=53.66667, MSE=580.6222
## left son=22812 (6 obs) right son=22813 (9 obs)
## Primary splits:
## Review Rating splits as --------R---L-RR--L-------, improve=0.43660060, (0 missing)
## Previous Purchases < 30.5 to the right, improve=0.13943720, (0 missing)
## Age < 47.5 to the right, improve=0.01617039, (0 missing)
## Surrogate splits:
## Previous Purchases < 22.5 to the left, agree=0.667, adj=0.167, (0 split)
##
## Node number 11407: 18 observations
## mean=70.27778, MSE=454.534
##
## Node number 11436: 7 observations
## mean=40.57143, MSE=381.102
##
## Node number 11437: 7 observations, complexity param=0.001198202
## mean=66.85714, MSE=527.551
## left son=22874 (3 obs) right son=22875 (4 obs)
## Primary splits:
## Previous Purchases < 12 to the right, improve=0.7645680, (0 missing)
## Age < 33.5 to the right, improve=0.2332721, (0 missing)
## Review Rating splits as -LL-----L---R-------------, improve=0.1549001, (0 missing)
## Surrogate splits:
## Review Rating splits as -RL-----R---R-------------, agree=0.714, adj=0.333, (0 split)
##
## Node number 12658: 9 observations
## mean=54.22222, MSE=330.6173
##
## Node number 12659: 7 observations
## mean=81.28571, MSE=305.6327
##
## Node number 22812: 6 observations
## mean=34.16667, MSE=123.8056
##
## Node number 22813: 9 observations
## mean=66.66667, MSE=462.6667
##
## Node number 22874: 3 observations
## mean=43.66667, MSE=140.2222
##
## Node number 22875: 4 observations
## mean=84.25, MSE=112.1875# Graficar el árbol
rpart.plot(modelo_arbol, main = "Árbol de Decisión")## Warning: labs do not fit even at cex 0.15, there may be some overplotting
# Conclusión
El árbol permite identificar perfiles de clientes de alto valor, agrupándolos según sus características. Es muy útil para segmentar el mercado y dirigir campañas específicas.
Por ejemplo, el árbol podría mostrar reglas como:
Clientes con edad menor a 30 y más de 2 compras previas tienden a gastar más.
Clientes con baja calificación en reseñas y pocas compras anteriores tienden a gastar menos.
Estas divisiones permiten identificar patrones de comportamiento de compra.
Conclusión del modelo
El modelo de árbol de decisión ha identificado relaciones entre características del cliente y el monto de compra, permitiendo hacer predicciones más informadas.
Es una herramienta intuitiva y visual, útil para comprender los factores que influyen en el gasto del cliente.
Si bien los árboles no son tan precisos como otros modelos más complejos (como random forests o regresión múltiple), ofrecen gran valor interpretativo, lo cual es ideal en contextos de análisis exploratorio y toma de decisiones de negocio.
Para mejorar el modelo, se puede:
Incluir más variables relevantes (como género, categoría del producto, dispositivo de compra).
Tratar valores faltantes o atípicos.
Validar el modelo con una muestra distinta o mediante validación cruzada.