Ejemplo cesta de la compra

Nombre : Noboa Andres

El algoritmo a priori se usa para encontrar conjuntos de elementos frecuentes en un conjunto de datos para la minería de reglas de asociación. Se llama a priori porque utiliza el conocimiento previo de las propiedades frecuentes de los conjuntos de elementos. Aplicamos un enfoque iterativo o una búsqueda por niveles en la que se utilizan conjuntos de elementos k frecuentes para encontrar conjuntos de elementos k+1. Para mejorar la eficiencia de la generación nivelada de conjuntos de elementos frecuentes, se utiliza una propiedad importante llamada propiedad Apriori que ayuda a reducir el espacio de búsqueda. Es muy fácil implementar este algoritmo utilizando el lenguaje de programación R.

library(readr)
library(arules)
library(magrittr)

library(tidyverse)

datos <- read_csv (file = "./datos_groceries.csv", col_names = TRUE,show_col_types = FALSE)
head(df,5)

transacciones <- read.transactions(file = "./datos_groceries.csv",
                                   format = "single",
                                   sep = ",",
                                   header = TRUE,
                                   cols = c("id_compra", "item"),
                                   rm.duplicates = TRUE)

transacciones

transactions in sparse format with
 9835 transactions (rows) and
 169 items (columns)

colnames(transacciones)[1:5]

[1] "abrasive cleaner" "artif. sweetener" "baby cosmetics"   "baby food"        "bags"            

rownames(transacciones)[1:5]

[1] "1"    "10"   "100"  "1000" "1001"

transacciones

transactions in sparse format with
 9835 transactions (rows) and
 169 items (columns)

datos_matriz <- df %>%
                as.data.frame() %>%
                mutate(valor = 1) %>%
                spread(key = item, value = valor, fill = 0) %>%
                column_to_rownames(var = "id_compra") %>%
                as.matrix()

transacciones <- as(datos_matriz, Class = "transactions")
transacciones

transactions in sparse format with
 9835 transactions (rows) and
 169 items (columns)

CONVERSIÓN DE UNA MATRIZ A UN OBJETO TIPO TRANSACTION

inspect(transacciones[1:5])

NA

df_transacciones <- as(transacciones, Class = "data.frame")
# Para que el tamaño de la tabla se ajuste mejor, se convierte el dataframe a tibble
as.tibble(df_transacciones) %>% head()

tamanyos <- size(transacciones)
summary(tamanyos)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   2.000   3.000   4.409   6.000  32.000 

data.frame(tamanyos) %>%
  ggplot(aes(x = tamanyos)) +
  geom_histogram() +
  labs(title = "Distribución del tamaño de las transacciones",
       x = "Tamaño") +
  theme_bw()

`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

frecuencia_items <- itemFrequency(x = transacciones, type = "relative")
frecuencia_items %>% sort(decreasing = TRUE) %>% head(5)

      whole milk other vegetables       rolls/buns             soda           yogurt 
       0.2555160        0.1934926        0.1839349        0.1743772        0.1395018 

frecuencia_items <- itemFrequency(x = transacciones, type = "absolute")
frecuencia_items %>% sort(decreasing = TRUE) %>% head(5)

      whole milk other vegetables       rolls/buns             soda           yogurt 
            2513             1903             1809             1715             1372 

soporte <- 30 / dim(transacciones)[1]
itemsets <- apriori(data = transacciones,
                    parameter = list(support = soporte,
                                     minlen = 1,
                                     maxlen = 20,
                                     target = "frequent itemset"))

Apriori

Parameter specification:

Algorithmic control:

Absolute minimum support count: 30 

set item appearances ...[0 item(s)] done [0.00s].
set transactions ...[169 item(s), 9835 transaction(s)] done [0.00s].
sorting and recoding items ... [136 item(s)] done [0.00s].
creating transaction tree ... done [0.00s].
checking subsets of size 1 2 3 4 5 done [0.00s].
sorting transactions ... done [0.00s].
writing ... [2226 set(s)] done [0.00s].
creating S4 object  ... done [0.00s].

summary(itemsets)

set of 2226 itemsets

most frequent items:
      whole milk other vegetables           yogurt  root vegetables       rolls/buns          (Other) 
             556              468              316              251              241             3536 

element (itemset/transaction) length distribution:sizes
   1    2    3    4    5 
 136 1140  850   98    2 

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   2.000   2.000   2.412   3.000   5.000 

summary of quality measures:
    support             count        
 Min.   :0.003050   Min.   :  30.00  
 1st Qu.:0.003660   1st Qu.:  36.00  
 Median :0.004779   Median :  47.00  
 Mean   :0.007879   Mean   :  77.49  
 3rd Qu.:0.007219   3rd Qu.:  71.00  
 Max.   :0.255516   Max.   :2513.00  

includes transaction ID lists: FALSE 

mining info:

# Se muestran los top 20 itemsets de mayor a menor soporte
top_20_itemsets <- sort(itemsets, by = "support", decreasing = TRUE)[1:20]
inspect(top_20_itemsets)

#Reglas de asociación
soporte <- 30 / dim(transacciones)[1]
reglas <- apriori(data = transacciones,
                  parameter = list(support = soporte,
                                   confidence = 0.70,
                                   # Se especifica que se creen reglas
                                   target = "rules"))

Apriori

Parameter specification:

Algorithmic control:

Absolute minimum support count: 30 

set item appearances ...[0 item(s)] done [0.00s].
set transactions ...[169 item(s), 9835 transaction(s)] done [0.00s].
sorting and recoding items ... [136 item(s)] done [0.00s].
creating transaction tree ... done [0.00s].
checking subsets of size 1 2 3 4 5 done [0.00s].
writing ... [19 rule(s)] done [0.00s].
creating S4 object  ... done [0.00s].

inspect(sort(x = reglas, decreasing = TRUE, by = "confidence"))

Confidencia Mayor e inferior

Para una confidencia mayor a 0.80

soporte <- 30 / dim(transacciones)[1]
reglas <- apriori(data = transacciones,
                  parameter = list(support = soporte,
                                   confidence = 0.80,
                                   # Se especifica que se creen reglas
                                   target = "rules"))

Apriori

Parameter specification:

Algorithmic control:

Absolute minimum support count: 30 

set item appearances ...[0 item(s)] done [0.00s].
set transactions ...[169 item(s), 9835 transaction(s)] done [0.00s].
sorting and recoding items ... [136 item(s)] done [0.00s].
creating transaction tree ... done [0.00s].
checking subsets of size 1 2 3 4 5 done [0.00s].
writing ... [1 rule(s)] done [0.00s].
creating S4 object  ... done [0.00s].

inspect(sort(x = reglas, decreasing = TRUE, by = "confidence"))

Para una confidencia menor a 0.30

soporte <- 30 / dim(transacciones)[1]
reglas <- apriori(data = transacciones,
                  parameter = list(support = soporte,
                                   confidence = 0.30,
                                   # Se especifica que se creen reglas
                                   target = "rules"))

Apriori

Parameter specification:

Algorithmic control:

Absolute minimum support count: 30 

set item appearances ...[0 item(s)] done [0.00s].
set transactions ...[169 item(s), 9835 transaction(s)] done [0.00s].
sorting and recoding items ... [136 item(s)] done [0.00s].
creating transaction tree ... done [0.00s].
checking subsets of size 1 2 3 4 5 done [0.01s].
writing ... [1361 rule(s)] done [0.00s].
creating S4 object  ... done [0.00s].

inspect(sort(x = reglas, decreasing = FALSE, by = "confidence"))