#Estimaciones de Modelos Espaciales, intento uno:
#Cargar la libreria spdep
library(spdep) # Econometria espacial
## Loading required package: sp
## Loading required package: Matrix
## Loading required package: spData
## To access larger datasets in this package, install the spDataLarge
## package with: `install.packages('spDataLarge',
## repos='https://nowosad.github.io/drat/', type='source'))`
library(maptools) # leer archivos shapfiles y elaborar mapas
## Checking rgeos availability: FALSE
## Note: when rgeos is not available, polygon geometry computations in maptools depend on gpclib,
## which has a restricted licence. It is disabled by default;
## to enable gpclib, type gpclibPermit()
library(RColorBrewer) # Eleccion de colores
library(classInt) # m??todos para clasificar
library(sf) #adicion
## Linking to GEOS 3.6.1, GDAL 2.2.3, proj.4 4.9.3
library(rgdal) #adicion
## rgdal: version: 1.2-20, (SVN revision 725)
## Geospatial Data Abstraction Library extensions to R successfully loaded
## Loaded GDAL runtime: GDAL 2.2.3, released 2017/11/20
## Path to GDAL shared files: C:/Users/Sergio Mora/Documents/R/win-library/3.5/rgdal/gdal
## GDAL binary built with GEOS: TRUE
## Loaded PROJ.4 runtime: Rel. 4.9.3, 15 August 2016, [PJ_VERSION: 493]
## Path to PROJ.4 shared files: C:/Users/Sergio Mora/Documents/R/win-library/3.5/rgdal/proj
## Linking to sp version: 1.2-7
#Cambiar el directorio de trabajo
setwd("C:/Users/Sergio Mora/Downloads")
#Leer archivos shapes y transformarlo en objeto Shape y DataFrame
empleo <- readOGR("Zona_Centro.shp")
## OGR data source with driver: ESRI Shapefile
## Source: "C:\Users\Sergio Mora\Downloads\Zona_Centro.shp", layer: "Zona_Centro"
## with 174 features
## It has 39 fields
## Integer64 fields read as strings: ID
summary(empleo)
## Object of class SpatialPolygonsDataFrame
## Coordinates:
## min max
## x -87.84584 -87.17459
## y 24.56797 25.16453
## Is projected: NA
## proj4string : [NA]
## Data attributes:
## ID CVEGEO NOM_ENT
## 266 : 1 09002 : 1 Distrito Federal: 16
## 267 : 1 09003 : 1 Morelos : 33
## 268 : 1 09004 : 1 MÚxico :125
## 269 : 1 09005 : 1
## 270 : 1 09006 : 1
## 271 : 1 09007 : 1
## (Other):168 (Other):168
## NOM_MUN POBTOT POBMAS
## Zacualpan : 2 Min. : 4051 Min. : 2012
## -lvaro Obreg¾n : 1 1st Qu.: 18469 1st Qu.: 9054
## Acambay : 1 Median : 44852 Median : 22188
## Acolman : 1 Mean : 148300 Mean : 71778
## Aculco : 1 3rd Qu.: 136118 3rd Qu.: 67958
## Almoloya de Alquisiras: 1 Max. :1815786 Max. :880998
## (Other) :167
## POBFEM PROM_HNV GRAPROES_M GRAPROES_F
## Min. : 2018 Min. :1.360 Min. : 5.610 Min. : 4.720
## 1st Qu.: 9407 1st Qu.:2.160 1st Qu.: 7.508 1st Qu.: 7.183
## Median : 22614 Median :2.360 Median : 8.515 Median : 8.175
## Mean : 76522 Mean :2.421 Mean : 8.514 Mean : 8.172
## 3rd Qu.: 68157 3rd Qu.:2.618 3rd Qu.: 9.385 3rd Qu.: 9.027
## Max. :934788 Max. :3.590 Max. :14.110 Max. :13.050
##
## PEA PEA_M PEA_F PE_INAC
## Min. : 1350 Min. : 1035 Min. : 189 Min. : 1405
## 1st Qu.: 6742 1st Qu.: 4962 1st Qu.: 1738 1st Qu.: 7071
## Median : 15462 Median : 11788 Median : 4514 Median : 16165
## Mean : 62670 Mean : 39849 Mean : 22821 Mean : 51245
## 3rd Qu.: 51654 3rd Qu.: 35943 3rd Qu.: 16036 3rd Qu.: 45507
## Max. :792297 Max. :491236 Max. :301061 Max. :626317
##
## PE_INAC_M PE_INAC_F POCUPADA POCUPADA_M
## Min. : 385 Min. : 1017 Min. : 1158 Min. : 953
## 1st Qu.: 1642 1st Qu.: 5380 1st Qu.: 6426 1st Qu.: 4643
## Median : 4012 Median : 12258 Median : 14470 Median : 10912
## Mean : 14365 Mean : 36880 Mean : 59568 Mean : 37554
## 3rd Qu.: 12075 3rd Qu.: 34270 3rd Qu.: 49053 3rd Qu.: 34220
## Max. :183133 Max. :443184 Max. :752268 Max. :463081
##
## POCUPADA_F PDESOCUP PDESOCUP_M PDESOCUP_F
## Min. : 183 Min. : 33.0 Min. : 28.0 Min. : 4.00
## 1st Qu.: 1708 1st Qu.: 313.2 1st Qu.: 263.8 1st Qu.: 41.25
## Median : 4406 Median : 860.0 Median : 739.0 Median : 129.00
## Mean : 22014 Mean : 3102.0 Mean : 2295.0 Mean : 807.03
## 3rd Qu.: 15687 3rd Qu.: 2642.5 3rd Qu.: 2103.0 3rd Qu.: 431.25
## Max. :289187 Max. :40029.0 Max. :28155.0 Max. :11874.00
##
## PSINDER PDER_SS PDER_IMSS PDER_ISTE
## Min. : 591 Min. : 1971 Min. : 43 Min. : 22.0
## 1st Qu.: 6758 1st Qu.: 11189 1st Qu.: 1311 1st Qu.: 453.5
## Median : 17835 Median : 26248 Median : 5560 Median : 1121.0
## Mean : 55903 Mean : 89534 Mean : 45848 Mean : 11061.9
## 3rd Qu.: 47760 3rd Qu.: 83030 3rd Qu.: 34660 3rd Qu.: 5621.8
## Max. :699848 Max. :1096323 Max. :593406 Max. :193469.0
##
## PDER_ISTEE PDER_SEGP GRAPROES N_CLASE_CR
## Min. : 1.0 Min. : 713 Min. : 5.150 Min. :1
## 1st Qu.: 222.2 1st Qu.: 6550 1st Qu.: 7.353 1st Qu.:1
## Median : 570.0 Median : 15484 Median : 8.395 Median :1
## Mean : 1851.9 Mean : 23533 Mean : 8.334 Mean :1
## 3rd Qu.: 1511.8 3rd Qu.: 30102 3rd Qu.: 9.217 3rd Qu.:1
## Max. :57934.0 Max. :274958 Max. :13.520 Max. :1
##
## ESCOLA_15 POB_15 CORE BOHEMIOS
## Min. : 3.130 Min. : 2613 Min. : 80 Min. : 0.00
## 1st Qu.: 5.270 1st Qu.: 12779 1st Qu.: 802 1st Qu.: 30.25
## Median : 6.165 Median : 29019 Median : 1854 Median : 107.00
## Mean : 6.373 Mean : 108714 Mean : 15905 Mean : 734.32
## 3rd Qu.: 7.008 3rd Qu.: 93237 3rd Qu.: 8286 3rd Qu.: 477.00
## Max. :12.680 Max. :1353286 Max. :209994 Max. :8220.00
##
## NO_CREATIV CORE_P NO_CREAT_P BOHE_P
## Min. : 28.0 Min. : 1.160 Min. : 0.410 Min. :0.0000
## 1st Qu.: 326.0 1st Qu.: 5.760 1st Qu.: 2.315 1st Qu.:0.2100
## Median : 955.5 Median : 8.230 Median : 4.180 Median :0.4050
## Mean : 9149.6 Mean : 9.492 Mean : 4.868 Mean :0.4764
## 3rd Qu.: 4809.0 3rd Qu.:11.670 3rd Qu.: 6.140 3rd Qu.:0.6575
## Max. :111629.0 Max. :38.260 Max. :21.080 Max. :1.9900
##
## OID
## Min. : 1.00
## 1st Qu.: 44.25
## Median : 87.50
## Mean : 87.50
## 3rd Qu.:130.75
## Max. :174.00
##
# Logaritmo del Empleo
lempleo <- log(empleo$POCUPADA)
# Capital Humano y logaritmo del capital humano, a??os de escolaridad
ch <- empleo$ESCOLA_15
lch <- log(empleo$ESCOLA_15)
# Caracteristicas estadisticas del empleo y capital humano
summary(empleo$POCUPADA)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1158 6426 14470 59568 49053 752268
summary(empleo$ESCOLA_15)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.130 5.270 6.165 6.373 7.008 12.680
summary(lempleo)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 7.054 8.768 9.580 9.882 10.801 13.531
summary(ch)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.130 5.270 6.165 6.373 7.008 12.680
summary(lch)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.141 1.662 1.819 1.821 1.947 2.540
class(empleo)
## [1] "SpatialPolygonsDataFrame"
## attr(,"package")
## [1] "sp"
# Construir lista de vecinos tipo Queen de poligonos
pr.nb <- poly2nb(empleo, queen=TRUE)
# Matriz de ponderacion W estandarizada
wqueen <- nb2listw(pr.nb, style="W")
# Caracteristicas de la Matriz W tipo Queen
summary(wqueen)
## Characteristics of weights list object:
## Neighbour list object:
## Number of regions: 174
## Number of nonzero links: 950
## Percentage nonzero weights: 3.137799
## Average number of links: 5.45977
## Link number distribution:
##
## 1 2 3 4 5 6 7 8 9 10 11 14
## 3 3 19 41 32 26 20 16 9 3 1 1
## 3 least connected regions:
## 76 94 120 with 1 link
## 1 most connected region:
## 116 with 14 links
##
## Weights style: W
## Weights constants summary:
## n nn S0 S1 S2
## W 174 30276 174 70.54671 724.6509
# Grafica con la conexicion espacial
cent <- coordinates(empleo)
plot(empleo, border="grey", lwd=1.5)
plot(pr.nb,cent, add=T, col="darkred")
# Estadistico de Moran
moran_lempleo <- moran.test(lempleo, wqueen,randomisation=TRUE, alternative="two.sided", na.action=na.exclude)
moran_ch <- moran.test(ch, wqueen,randomisation=TRUE, alternative="two.sided", na.action=na.exclude)
moran_lch <- moran.test(lch, wqueen,randomisation=TRUE, alternative="two.sided", na.action=na.exclude)
#Ver resultados
print(moran_lempleo)
##
## Moran I test under randomisation
##
## data: lempleo
## weights: wqueen
##
## Moran I statistic standard deviate = 12.466, p-value < 2.2e-16
## alternative hypothesis: two.sided
## sample estimates:
## Moran I statistic Expectation Variance
## 0.587496960 -0.005780347 0.002264968
print(moran_ch)
##
## Moran I test under randomisation
##
## data: ch
## weights: wqueen
##
## Moran I statistic standard deviate = 14.652, p-value < 2.2e-16
## alternative hypothesis: two.sided
## sample estimates:
## Moran I statistic Expectation Variance
## 0.687457813 -0.005780347 0.002238704
print(moran_lch)
##
## Moran I test under randomisation
##
## data: lch
## weights: wqueen
##
## Moran I statistic standard deviate = 14.843, p-value < 2.2e-16
## alternative hypothesis: two.sided
## sample estimates:
## Moran I statistic Expectation Variance
## 0.698986435 -0.005780347 0.002254493
# Grafica de diagrama de dispersion de Moran
moran.plot(lempleo, wqueen, pch=20)
moran.plot(ch, wqueen, pch=20)
moran.plot(lch, wqueen, pch=20)
#Clasificador con quintiles y colores especificos
# Mapa de quintiles del logaritmo del empleo
brks <- round(quantile(lempleo, probs=seq(0,1,0.25)), digits=2)
colours <- brewer.pal(4,"Reds")
plot(empleo, col=colours[findInterval(lempleo, brks, all.inside=TRUE)],
axes=F)
legend(x=-87.9, y=25.2, legend=leglabs(brks), fill=colours, bty="n")
#invisible(title(main=paste("EMPLEO", sep="\n")))
#box()
# Mapa de quintiles del logaritmo del capital humano
brks <- round(quantile(lch, probs=seq(0,1,0.25)), digits=2)
colours <- brewer.pal(4,"Blues")
plot(empleo, col=colours[findInterval(lch, brks, all.inside=TRUE)],
axes=F)
legend(x=-87.9, y=25.2, legend=leglabs(brks), fill=colours, bty="n")
#invisible(title(main=paste("CAPITAL HUMANO", sep="\n")))
#box()
# An??lisis LISA
# Valores de referencia z de la distribuci??n t
z <- c(1.65, 1.96)
zc <- c(2.8284, 3.0471)
# Estimaci??n de indice de Moran local (Ii)
f.Ii <- localmoran(lempleo, wqueen)
zIi <- f.Ii[,"Z.Ii"] # Asignaci??n de la distribuci??n Z del Ii
mx <- max(zIi)
mn <- min(zIi)
# Mapa de signficancia para los z-scores
pal <- c("white", "green", "darkgreen")
z3.Ii <- classIntervals(zIi, n=3, style="fixed", fixedBreaks=c(mn, z, mx))
cols.Ii <- findColours(z3.Ii, pal)
plot(empleo, col=cols.Ii)
brks <- round(z3.Ii$brks,4)
leg <- paste(brks[-4], brks[-1], sep=" - ")
legend(x=-87.9, y=25.2, fill=pal, legend=leg, bty="n")
# Mapa de los grupos de cluster
pal.rb <- c("skyblue1","white","blue","red")
z4.Ii <- classIntervals(zIi, n=4, style="fixed", fixedBreaks=c(min(f.Ii), -z[1], z[1], z[2], max(f.Ii)))
cols.Ii <- findColours(z4.Ii, pal.rb)
plot(empleo, col=cols.Ii)
brks <- round(z4.Ii$brks,4)
leg <- paste(brks[-5], brks[-1], sep=" - ")
legend(x=-87.9, y=25.2, fill= pal.rb, legend=leg, bty="n")
# Modelo OLS
ModeloEmpleo_OLS <- lm(lempleo ~ lch , data=empleo)
summary(ModeloEmpleo_OLS)
##
## Call:
## lm(formula = lempleo ~ lch, data = empleo)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.3948 -0.8243 0.0286 0.7611 2.8910
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.8775 0.6763 5.734 4.33e-08 ***
## lch 3.2969 0.3680 8.960 5.26e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.2 on 172 degrees of freedom
## Multiple R-squared: 0.3182, Adjusted R-squared: 0.3142
## F-statistic: 80.27 on 1 and 172 DF, p-value: 5.264e-16
# Prueba de Moran a residuales del modelo OLS
I_Moran <- lm.morantest(ModeloEmpleo_OLS,wqueen)
print(I_Moran)
##
## Global Moran I for regression residuals
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## Moran I statistic standard deviate = 8.0244, p-value = 5.1e-16
## alternative hypothesis: greater
## sample estimates:
## Observed Moran I Expectation Variance
## 0.368395591 -0.009877828 0.002222204
#Pruebas de Multiplicadores de Lagranges lm.LMtests(columbus.lm,col.listw,test=c("LMerr","RLMerr","LMlag","RLMlag","SARMA"))
lm.LMtests(ModeloEmpleo_OLS,wqueen,test=c("LMerr","RLMerr","LMlag","RLMlag","SARMA"))
##
## Lagrange multiplier diagnostics for spatial dependence
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## LMerr = 58.244, df = 1, p-value = 2.32e-14
##
##
## Lagrange multiplier diagnostics for spatial dependence
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## RLMerr = 0.0032553, df = 1, p-value = 0.9545
##
##
## Lagrange multiplier diagnostics for spatial dependence
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## LMlag = 70.722, df = 1, p-value < 2.2e-16
##
##
## Lagrange multiplier diagnostics for spatial dependence
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## RLMlag = 12.482, df = 1, p-value = 0.0004109
##
##
## Lagrange multiplier diagnostics for spatial dependence
##
## data:
## model: lm(formula = lempleo ~ lch, data = empleo)
## weights: wqueen
##
## SARMA = 70.726, df = 2, p-value = 4.441e-16
# Modelos Espaciales
# Estimar el Modelo Rezago Espacial
ModeloEmpleo_lag <- lagsarlm(lempleo ~ lch , data=empleo,wqueen)
summary(ModeloEmpleo_lag)
##
## Call:lagsarlm(formula = lempleo ~ lch, data = empleo, listw = wqueen)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.1783841 -0.5655122 -0.0045741 0.5698724 2.4373878
##
## Type: lag
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.73366 0.61410 1.1947 0.2322
## lch 1.32893 0.32970 4.0307 5.562e-05
##
## Rho: 0.67009, LR test value: 65.113, p-value: 6.6613e-16
## Asymptotic standard error: 0.06377
## z-value: 10.508, p-value: < 2.22e-16
## Wald statistic: 110.42, p-value: < 2.22e-16
##
## Log likelihood: -245.0376 for lag model
## ML residual variance (sigma squared): 0.87624, (sigma: 0.93607)
## Number of observations: 174
## Number of parameters estimated: 4
## AIC: 498.08, (AIC for lm: 561.19)
## LM test for residual autocorrelation
## test value: 0.11791, p-value: 0.73132
# Estimar el modelo de Error Espacial
ModeloEmpleo_err <- errorsarlm(lempleo ~ lch , data=empleo,wqueen)
summary(ModeloEmpleo_err)
##
## Call:errorsarlm(formula = lempleo ~ lch, data = empleo, listw = wqueen)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.229495 -0.603003 0.050277 0.613023 2.550370
##
## Type: error
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 6.47006 0.94208 6.8679 6.516e-12
## lch 1.64366 0.49579 3.3152 0.0009157
##
## Lambda: 0.71817, LR test value: 58.604, p-value: 1.9318e-14
## Asymptotic standard error: 0.061095
## z-value: 11.755, p-value: < 2.22e-16
## Wald statistic: 138.18, p-value: < 2.22e-16
##
## Log likelihood: -248.2923 for error model
## ML residual variance (sigma squared): 0.89031, (sigma: 0.94356)
## Number of observations: 174
## Number of parameters estimated: 4
## AIC: 504.58, (AIC for lm: 561.19)
# Estimar modelo SARAR
ModeloEmpleo_sarar <- sacsarlm(lempleo ~ lch , data=empleo, wqueen, type="sac")
summary(ModeloEmpleo_sarar)
##
## Call:sacsarlm(formula = lempleo ~ lch, data = empleo, listw = wqueen,
## type = "sac")
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.142804 -0.581190 -0.016665 0.575159 2.415785
##
## Type: sac
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.56642 0.74371 0.7616 0.446294
## lch 1.24960 0.41976 2.9770 0.002911
##
## Rho: 0.70235
## Asymptotic standard error: 0.11906
## z-value: 5.899, p-value: 3.6575e-09
## Lambda: -0.073671
## Asymptotic standard error: 0.24171
## z-value: -0.30479, p-value: 0.76053
##
## LR test value: 65.214, p-value: 6.8834e-15
##
## Log likelihood: -244.9869 for sac model
## ML residual variance (sigma squared): 0.86265, (sigma: 0.92879)
## Number of observations: 174
## Number of parameters estimated: 5
## AIC: 499.97, (AIC for lm: 561.19)
#Estimar el modelo de Durbin Rezago Espacial
ModeloEmpleo_lag_durbin <- lagsarlm(lempleo ~ lch , data=empleo,wqueen, type="mixed")
summary(ModeloEmpleo_lag_durbin)
##
## Call:lagsarlm(formula = lempleo ~ lch, data = empleo, listw = wqueen,
## type = "mixed")
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.059210 -0.599797 -0.023718 0.585009 2.417886
##
## Type: mixed
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.54567 0.67171 0.8124 0.41659
## lch 1.03039 0.55554 1.8547 0.06363
## lag.lch 0.50523 0.71928 0.7024 0.48242
##
## Rho: 0.6515, LR test value: 52.636, p-value: 4.0157e-13
## Asymptotic standard error: 0.06901
## z-value: 9.4407, p-value: < 2.22e-16
## Wald statistic: 89.127, p-value: < 2.22e-16
##
## Log likelihood: -244.8002 for mixed model
## ML residual variance (sigma squared): 0.88042, (sigma: 0.93831)
## Number of observations: 174
## Number of parameters estimated: 5
## AIC: 499.6, (AIC for lm: 550.24)
## LM test for residual autocorrelation
## test value: 0.024694, p-value: 0.87513
#Estimar el modelo de Durbin Error Espacial
ModeloEmpleo_err_durbin <- errorsarlm(lempleo ~ lch , data=empleo,wqueen, etype="emixed")
summary(ModeloEmpleo_err_durbin)
##
## Call:errorsarlm(formula = lempleo ~ lch, data = empleo, listw = wqueen,
## etype = "emixed")
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.127211 -0.570143 -0.040392 0.575190 2.423206
##
## Type: error
## Coefficients: (asymptotic standard errors)
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.62363 1.52942 1.7154 0.086264
## lch 1.35737 0.50232 2.7022 0.006888
## lag.lch 2.47167 0.86244 2.8659 0.004158
##
## Lambda: 0.65455, LR test value: 52.865, p-value: 3.5738e-13
## Asymptotic standard error: 0.069289
## z-value: 9.4467, p-value: < 2.22e-16
## Wald statistic: 89.24, p-value: < 2.22e-16
##
## Log likelihood: -244.6857 for error model
## ML residual variance (sigma squared): 0.87821, (sigma: 0.93713)
## Number of observations: 174
## Number of parameters estimated: 5
## AIC: 499.37, (AIC for lm: 550.24)
#Comparar modelos de error espacial y durbin error espacial(Hipoteis nula son iguales), con pruebas de Razon de Verosimilitud
LR.sarlm(ModeloEmpleo_err_durbin, ModeloEmpleo_err)
##
## Likelihood ratio for spatial linear models
##
## data:
## Likelihood ratio = 7.2132, df = 1, p-value = 0.007237
## sample estimates:
## Log likelihood of ModeloEmpleo_err_durbin
## -244.6857
## Log likelihood of ModeloEmpleo_err
## -248.2923
#Comparar modelos de rezago espacial y durbin rezago espacial(Hipoteis nula son iguales), con pruebas de Razon de Verosimilitud
LR.sarlm(ModeloEmpleo_lag_durbin, ModeloEmpleo_lag)
##
## Likelihood ratio for spatial linear models
##
## data:
## Likelihood ratio = 0.47491, df = 1, p-value = 0.4907
## sample estimates:
## Log likelihood of ModeloEmpleo_lag_durbin
## -244.8002
## Log likelihood of ModeloEmpleo_lag
## -245.0376
#Probar la hipotesis de factor com?n
durbin.test1 <- LR.sarlm(ModeloEmpleo_lag_durbin,ModeloEmpleo_err)
print(durbin.test1)
##
## Likelihood ratio for spatial linear models
##
## data:
## Likelihood ratio = 6.9842, df = 1, p-value = 0.008223
## sample estimates:
## Log likelihood of ModeloEmpleo_lag_durbin
## -244.8002
## Log likelihood of ModeloEmpleo_err
## -248.2923
1 - pchisq(durbin.test1[[1]][1],2)
## Likelihood ratio
## 0.0304362
#Comparar modeloss con ANOVA
anova(ModeloEmpleo_lag, ModeloEmpleo_lag_durbin)
## Model df AIC logLik Test L.Ratio p-value
## ModeloEmpleo_lag 1 4 498.08 -245.04 1
## ModeloEmpleo_lag_durbin 2 5 499.60 -244.80 2 0.47491 0.49074
#Breusch-Pagan test for spatial models
bptest.sarlm(ModeloEmpleo_lag, studentize=FALSE)
##
## Breusch-Pagan test
##
## data:
## BP = 6.5471, df = 1, p-value = 0.01051
# Evaluar impactos en modelos de Rezago Espacial y Durbidn Rezago Espacial
#Matrices para el calculo
W <- as(as_dgRMatrix_listw(wqueen), "CsparseMatrix")
trMatc <- trW(W, type="mult")
trMC <- trW(W, type="MC")
#Impactos modelo rezago espacial
impacts(ModeloEmpleo_lag, listw=wqueen)
## Impact measures (lag, exact):
## Direct Indirect Total
## lch 1.513396 2.514804 4.0282
impacts(ModeloEmpleo_lag, tr=trMatc)
## Impact measures (lag, trace):
## Direct Indirect Total
## lch 1.513396 2.51478 4.028176
impacts(ModeloEmpleo_lag, tr=trMC)
## Impact measures (lag, trace):
## Direct Indirect Total
## lch 1.51193 2.516246 4.028176
summary(impacts(ModeloEmpleo_lag, tr=trMatc, R=1000), zstats=TRUE, short=TRUE)
## Impact measures (lag, trace):
## Direct Indirect Total
## lch 1.513396 2.51478 4.028176
## ========================================================
## Simulation results (asymptotic variance matrix):
## ========================================================
## Simulated z-values:
## Direct Indirect Total
## lch 4.314809 3.32023 4.056422
##
## Simulated p-values:
## Direct Indirect Total
## lch 1.5974e-05 0.00089943 4.983e-05
summary(impacts(ModeloEmpleo_lag, tr=trMatc, R=1000, Q=10), zstats=TRUE, short=TRUE,reportQ=TRUE)
## Impact measures (lag, trace):
## Direct Indirect Total
## lch 1.513396 2.51478 4.028176
## =================================
## Impact components
## $direct
## lch
## Q1 1.328928126
## Q2 0.000000000
## Q3 0.111150166
## Q4 0.024136769
## Q5 0.022837893
## Q6 0.009817296
## Q7 0.006782716
## Q8 0.003663321
## Q9 0.002352407
## Q10 0.001386173
##
## $indirect
## lch
## Q1 0.00000000
## Q2 0.89050654
## Q3 0.48557277
## Q4 0.37572359
## Q5 0.24510606
## Q6 0.16973030
## Q7 0.11353102
## Q8 0.07695817
## Q9 0.05167156
## Q10 0.03481495
##
## $total
## lch
## Q1 1.32892813
## Q2 0.89050654
## Q3 0.59672293
## Q4 0.39986035
## Q5 0.26794396
## Q6 0.17954759
## Q7 0.12031373
## Q8 0.08062149
## Q9 0.05402396
## Q10 0.03620112
##
## ========================================================
## Simulation results (asymptotic variance matrix):
## ========================================================
## Simulated z-values:
## Direct Indirect Total
## lch 4.495309 3.49212 4.339877
##
## Simulated p-values:
## Direct Indirect Total
## lch 6.9469e-06 0.0004792 1.4256e-05
## ========================================================
## Simulated impact components z-values:
## $Direct
## lch
## Q1 4.216556
## Q2 NaN
## Q3 4.403352
## Q4 3.636735
## Q5 2.941886
## Q6 2.416673
## Q7 2.027602
## Q8 1.733148
## Q9 1.503590
## Q10 1.319433
##
## $Indirect
## lch
## Q1 NaN
## Q2 4.743845
## Q3 4.403352
## Q4 3.636735
## Q5 2.941886
## Q6 2.416673
## Q7 2.027602
## Q8 1.733148
## Q9 1.503590
## Q10 1.319433
##
## $Total
## lch
## Q1 4.216556
## Q2 4.743845
## Q3 4.403352
## Q4 3.636735
## Q5 2.941886
## Q6 2.416673
## Q7 2.027602
## Q8 1.733148
## Q9 1.503590
## Q10 1.319433
##
##
## Simulated impact components p-values:
## $Direct
## lch
## Q1 2.4806e-05
## Q2 NA
## Q3 1.0659e-05
## Q4 0.00027612
## Q5 0.00326220
## Q6 0.01566307
## Q7 0.04260087
## Q8 0.08306934
## Q9 0.13268699
## Q10 0.18702445
##
## $Indirect
## lch
## Q1 NA
## Q2 2.0970e-06
## Q3 1.0659e-05
## Q4 0.00027612
## Q5 0.00326220
## Q6 0.01566307
## Q7 0.04260087
## Q8 0.08306934
## Q9 0.13268699
## Q10 0.18702445
##
## $Total
## lch
## Q1 2.4806e-05
## Q2 2.0970e-06
## Q3 1.0659e-05
## Q4 0.00027612
## Q5 0.00326220
## Q6 0.01566307
## Q7 0.04260087
## Q8 0.08306934
## Q9 0.13268699
## Q10 0.18702445
#Impactos modelo Durbin Rezago Espacial
impacts(ModeloEmpleo_lag_durbin, listw=wqueen)
## Impact measures (mixed, exact):
## Direct Indirect Total
## lch 1.260805 3.145558 4.406363
impacto2 <- impacts(ModeloEmpleo_lag_durbin, tr=trMatc,R=200,Q=5)
summary(impacto2)
## Impact measures (mixed, trace):
## Direct Indirect Total
## lch 1.260805 3.145547 4.406352
## ========================================================
## Simulation results (asymptotic variance matrix):
## Direct:
##
## Iterations = 1:200
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 200
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## lch 1.34 0.4781 0.0338 0.0338
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## lch 0.2842 1.063 1.371 1.689 2.106
##
## ========================================================
## Indirect:
##
## Iterations = 1:200
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 200
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## lch 3.052 1.148 0.0812 0.0812
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## lch 0.8779 2.348 2.957 3.731 5.295
##
## ========================================================
## Total:
##
## Iterations = 1:200
## Thinning interval = 1
## Number of chains = 1
## Sample size per chain = 200
##
## 1. Empirical mean and standard deviation for each variable,
## plus standard error of the mean:
##
## Mean SD Naive SE Time-series SE
## lch 4.392 1.094 0.07735 0.07735
##
## 2. Quantiles for each variable:
##
## 2.5% 25% 50% 75% 97.5%
## lch 2.291 3.76 4.318 5.024 6.658
impacts(ModeloEmpleo_lag_durbin, tr=trMC)
## Impact measures (mixed, trace):
## Direct Indirect Total
## lch 1.259124 3.147228 4.406352
summary(impacto2, zstats=TRUE, short=TRUE, reportQ=TRUE)
## Impact measures (mixed, trace):
## Direct Indirect Total
## lch 1.260805 3.145547 4.406352
## =================================
## Impact components
## $direct
## lch
## Q1 1.03038525
## Q2 0.06131180
## Q3 0.09440862
## Q4 0.02910761
## Q5 0.02079919
##
## $indirect
## lch
## Q1 0.5052339
## Q2 0.9391435
## Q3 0.5573877
## Q4 0.3955374
## Q5 0.2558569
##
## $total
## lch
## Q1 1.5356192
## Q2 1.0004553
## Q3 0.6517963
## Q4 0.4246450
## Q5 0.2766561
##
## ========================================================
## Simulation results (asymptotic variance matrix):
## ========================================================
## Simulated z-values:
## Direct Indirect Total
## lch 2.803908 2.657586 4.015445
##
## Simulated p-values:
## Direct Indirect Total
## lch 0.0050487 0.0078702 5.9334e-05
## ========================================================
## Simulated impact components z-values:
## $Direct
## lch
## Q1 2.1937551
## Q2 0.5399261
## Q3 2.8388334
## Q4 2.4968814
## Q5 2.4867013
##
## $Indirect
## lch
## Q1 0.5922279
## Q2 4.6961090
## Q3 3.4074928
## Q4 3.3543396
## Q5 2.6838827
##
## $Total
## lch
## Q1 3.405669
## Q2 3.998737
## Q3 3.904515
## Q4 3.316421
## Q5 2.706860
##
##
## Simulated impact components p-values:
## $Direct
## lch
## Q1 0.0282530
## Q2 0.5892480
## Q3 0.0045279
## Q4 0.0125291
## Q5 0.0128934
##
## $Indirect
## lch
## Q1 0.55369801
## Q2 2.6516e-06
## Q3 0.00065563
## Q4 0.00079555
## Q5 0.00727726
##
## $Total
## lch
## Q1 0.00066002
## Q2 6.3681e-05
## Q3 9.4415e-05
## Q4 0.00091178
## Q5 0.00679228
summary(impacts(ModeloEmpleo_lag_durbin, tr=trMatc, R=200, Q=5), zstats=TRUE, short=TRUE,reportQ=TRUE)
## Impact measures (mixed, trace):
## Direct Indirect Total
## lch 1.260805 3.145547 4.406352
## =================================
## Impact components
## $direct
## lch
## Q1 1.03038525
## Q2 0.06131180
## Q3 0.09440862
## Q4 0.02910761
## Q5 0.02079919
##
## $indirect
## lch
## Q1 0.5052339
## Q2 0.9391435
## Q3 0.5573877
## Q4 0.3955374
## Q5 0.2558569
##
## $total
## lch
## Q1 1.5356192
## Q2 1.0004553
## Q3 0.6517963
## Q4 0.4246450
## Q5 0.2766561
##
## ========================================================
## Simulation results (asymptotic variance matrix):
## ========================================================
## Simulated z-values:
## Direct Indirect Total
## lch 2.290723 2.634592 4.064209
##
## Simulated p-values:
## Direct Indirect Total
## lch 0.021979 0.0084238 4.8196e-05
## ========================================================
## Simulated impact components z-values:
## $Direct
## lch
## Q1 1.7294140
## Q2 0.6402196
## Q3 2.7467847
## Q4 2.6329493
## Q5 2.8223840
##
## $Indirect
## lch
## Q1 0.6605046
## Q2 4.5906632
## Q3 3.5094829
## Q4 3.5905733
## Q5 2.8912818
##
## $Total
## lch
## Q1 3.451879
## Q2 4.018689
## Q3 4.037380
## Q4 3.539447
## Q5 2.926229
##
##
## Simulated impact components p-values:
## $Direct
## lch
## Q1 0.0837350
## Q2 0.5220298
## Q3 0.0060183
## Q4 0.0084647
## Q5 0.0047668
##
## $Indirect
## lch
## Q1 0.50893010
## Q2 4.4184e-06
## Q3 0.00044898
## Q4 0.00032995
## Q5 0.00383674
##
## $Total
## lch
## Q1 0.00055670
## Q2 5.8523e-05
## Q3 5.4052e-05
## Q4 0.00040097
## Q5 0.00343098
#basado en:
#Econometria Espacial, 2015
#Profesor: Miguel Angel Mendoza, FE-UNAM