Data Science and Wine Quality

 
 # Descision Tree
 wine <- read.csv('winequality-white.csv', sep=';', header = TRUE)
 wine <- rbind(wine, read.csv('winequality-red.csv', sep=';', header = TRUE))
 hist(wine$quality)
 
 # The wine quality values appear to follow a fairly normal, 
 #bell-shaped distribution, centered around a value of six.
 

 summary(wine)
 
 #partition the data
 wine_train <- wine[1:3248, ]
 wine_test <- wine[3249:6497, ]
 
 m.rpart <- rpart::rpart(quality ~ ., data = wine_train)
 m.rpart
 
 summary(m.rpart)
 ## Call:
 ## rpart::rpart(formula = quality ~ ., data = wine_train)
 ##   n= 3248 
 ## 
 ##           CP nsplit rel error    xerror       xstd
 ## 1 0.17309330      0 1.0000000 1.0009756 0.02583728
 ## 2 0.04673914      1 0.8269067 0.8284360 0.02410107
 ## 3 0.03098582      2 0.7801676 0.7908229 0.02375595
 ## 4 0.01380180      3 0.7491817 0.7603411 0.02229241
 ## 5 0.01113703      4 0.7353799 0.7515058 0.02204882
 ## 6 0.01000000      5 0.7242429 0.7434067 0.02183699
 ## 
 ## Variable importance
 ##              alcohol              density     volatile.acidity 
 ##                   39                   22                   13 
 ##            chlorides  free.sulfur.dioxide total.sulfur.dioxide 
 ##                   11                    6                    5 
 ##          citric.acid                   pH            sulphates 
 ##                    1                    1                    1 
 ##        fixed.acidity 
 ##                    1 
 ## 

```r
 library(rpart.plot)
 
 ## Loading required package: rpart
 
 rpart.plot(m.rpart, digits = 3)