library(car) #> Loading required package: carData library(readr)

library(readr)
partners <- read_delim("https://themys.sid.uncu.edu.ar/rpalma/R-cran/SCM_Perform/scm_perform.csv", 
    delim = ";", escape_double = FALSE, col_types = cols(Performance = col_factor(levels = c("good", 
        "fair", "poor"))), locale = locale(decimal_mark = ",", 
        grouping_mark = "."), trim_ws = TRUE)
summary(partners)
##  Inventory.Turnover   Lead.Time     Perfect.Order   Trnasport.Km/Ton
##  Min.   :4.300      Min.   :2.000   Min.   :1.000   Min.   :0.100   
##  1st Qu.:5.100      1st Qu.:2.800   1st Qu.:1.600   1st Qu.:0.300   
##  Median :5.800      Median :3.000   Median :4.350   Median :1.300   
##  Mean   :5.843      Mean   :3.057   Mean   :3.758   Mean   :1.199   
##  3rd Qu.:6.400      3rd Qu.:3.300   3rd Qu.:5.100   3rd Qu.:1.800   
##  Max.   :7.900      Max.   :4.400   Max.   :6.900   Max.   :2.500   
##  Performance
##  good:50    
##  fair:50    
##  poor:50    
##             
##             
##

Matriz de covarianza

library(scatterPlotMatrix)
scatterPlotMatrix(partners)

Entrenamiento de árbol de decisión

Esta técnica utiliza un set de datos representativos de una situaci’on y utilizando recursivamente el teoréma de Bayes puede armar un pronosticador o clasificador de datos. Es una t’ecnica parecida a la de clustering, pero m’as refinada, pues no se basa en reglas sino en parendizaje del set de datos usado como entrenamento. En el paquete party existen dos funciones ctree que se utiliza para entrenar y predict que se usa para pronosticar o generar la regla de decici’on que debemos usar.

library(party)
## Loading required package: grid
## Loading required package: mvtnorm
## Loading required package: modeltools
## Loading required package: stats4
## Loading required package: strucchange
## Loading required package: zoo
## 
## Attaching package: 'zoo'
## The following objects are masked from 'package:base':
## 
##     as.Date, as.Date.numeric
## Loading required package: sandwich
attach(partners)
str(partners)    
## spc_tbl_ [150 × 5] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
##  $ Inventory.Turnover: num [1:150] 5.1 4.9 4.7 4.6 5 5.4 4.6 5 4.4 4.9 ...
##  $ Lead.Time         : num [1:150] 3.5 3 3.2 3.1 3.6 3.9 3.4 3.4 2.9 3.1 ...
##  $ Perfect.Order     : num [1:150] 1.4 1.4 1.3 1.5 1.4 1.7 1.4 1.5 1.4 1.5 ...
##  $ Trnasport.Km/Ton  : num [1:150] 0.2 0.2 0.2 0.2 0.2 0.4 0.3 0.2 0.2 0.1 ...
##  $ Performance       : Factor w/ 3 levels "good","fair",..: 1 1 1 1 1 1 1 1 1 1 ...
##  - attr(*, "spec")=
##   .. cols(
##   ..   Inventory.Turnover = col_double(),
##   ..   Lead.Time = col_double(),
##   ..   Perfect.Order = col_double(),
##   ..   `Trnasport.Km/Ton` = col_double(),
##   ..   Performance = col_factor(levels = c("good", "fair", "poor"), ordered = FALSE, include_na = FALSE)
##   .. )
##  - attr(*, "problems")=<externalptr>
ind <- sample(2, nrow(partners), replace=TRUE, prob=c(0.7, 0.3))  
# toma una muestra  
ind  
##   [1] 2 1 1 2 1 2 2 2 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2
##  [38] 1 1 1 1 1 1 2 1 1 2 1 2 1 1 2 1 1 2 2 1 1 2 2 1 2 2 1 2 1 2 2 1 2 1 2 1 1
##  [75] 1 1 2 2 2 2 2 1 1 2 1 1 1 2 1 2 1 2 2 2 2 2 1 2 1 1 1 1 1 1 2 1 2 2 1 1 1
## [112] 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2 1 1 2 1 1 1 1 1
## [149] 1 1
# nos imprime la muestra tomada.

Entrenamiento con primer muestra

trainData <- partners [ind==1,]    
# genero un set de entrenamiento 
testData <- partners [ind==2,]    
# genero un set de datos de prueba
myFormula <- Performance ~ Inventory.Turnover + Lead.Time + `Trnasport.Km/Ton` + Perfect.Order 
transit_ctree <- ctree(myFormula, data=trainData)    
# creo el motor de entrenamiento
# Verificar las prediciones 
table(predict(transit_ctree), trainData$Performance) 
##       
##        good fair poor
##   good   40    0    0
##   fair    0   23    3
##   poor    0    1   39
print(transit_ctree) 
## 
##   Conditional inference tree with 4 terminal nodes
## 
## Response:  Performance 
## Inputs:  Inventory.Turnover, Lead.Time, Trnasport.Km/Ton, Perfect.Order 
## Number of observations:  106 
## 
## 1) Perfect.Order <= 1.9; criterion = 1, statistic = 99.592
##   2)*  weights = 40 
## 1) Perfect.Order > 1.9
##   3) Trnasport.Km/Ton <= 1.6; criterion = 1, statistic = 41.909
##     4) Perfect.Order <= 4.5; criterion = 0.993, statistic = 9.723
##       5)*  weights = 16 
##     4) Perfect.Order > 4.5
##       6)*  weights = 10 
##   3) Trnasport.Km/Ton > 1.6
##     7)*  weights = 40

Muestra el arbol entrenado vs el resto de la muestra

Prediccion con el dataset de prueba

table(predict(transit_ctree), trainData$Performance)
##       
##        good fair poor
##   good   40    0    0
##   fair    0   23    3
##   poor    0    1   39

Impresión del arbol de decision

plot(transit_ctree, las=2)