Re: 문제216. 와인 데이터를 Riper 알고리즘으로 분류하세요.(답글로 코드 올려주세요)

[12기_권세원]

wine <- read.csv("wine.csv", header= T, stringsAsFactors = T)
summary(wine)
str(wine)
dim(wine)

temp<- c()
temp2 <-c()
i<-0
repeat { 
  i<- i+1
set.seed(i) 
train_cnt <- round( 0.80 * dim(wine)[1])
train_index <- sample(1:dim(wine)[1], train_cnt, replace=F)

wine_train <- wine[train_index,  ]
wine_test  <- wine[-train_index, ]
nrow(wine_train)

model3 <-  JRip(Type~ ., data=wine_train)
model3

summary(model3)  

result3 <- predict( model3, wine_test[   , -1] )
library(gmodels)
g2 <- CrossTable( wine_test[ , 1],  result3) 
j <- sum(g2$prop.tbl*diag(3))

print(c(i,j))
if(j>=0.98) # 목표 정확도 설정
{print(c(i,j)) 
break }
else if( i==1200) # 목표 정확도에 도달하지 못했다면 최고값
{result <- data.frame( 'seed'= temp2, '정확도'= temp)
print(result[result$정확도==max(result$정확도),])
break}
}

wine <- read.csv("wine.csv", header= T, stringsAsFactors = T)
summary(wine)
str(wine)
dim(wine)


set.seed(4) 
train_cnt <- round( 0.80 * dim(wine)[1])
train_index <- sample(1:dim(wine)[1], train_cnt, replace=F)
  
wine_train <- wine[train_index,  ]
wine_test  <- wine[-train_index, ]
nrow(wine_train)
  
model3 <-  JRip(Type~ ., data=wine_train)
model3
  
summary(model3)  
  
result3 <- predict( model3, wine_test[   , -1] )
library(gmodels)
g2 <- CrossTable( wine_test[ , 1],  result3) 
j <- (g2$prop.tbl*diag(3))
print(j)
print(sum(j))

    y
x           t1        t2        t3
  t1 0.4166667 0.0000000 0.0000000
  t2 0.0000000 0.3611111 0.0000000
  t3 0.0000000 0.0000000 0.2222222

[1] 1 # 정확도