R Notebook

library(randomForest)
library(ggplot2)
library(dplyr)
library(stringr)
library(kernlab)
library(class)

red_wine <- read.csv("winequality-red.csv",sep = ";")
white_wine <- read.csv("winequality-white.csv",sep = ";")
bank <- read.csv("bank.csv", sep = ";")

print("Summary of red_wine and white_wine")

[1] "Summary of red_wine and white_wine"

summary(red_wine)

 fixed.acidity   volatile.acidity
 Min.   : 4.60   Min.   :0.1200  
 1st Qu.: 7.10   1st Qu.:0.3900  
 Median : 7.90   Median :0.5200  
 Mean   : 8.32   Mean   :0.5278  
 3rd Qu.: 9.20   3rd Qu.:0.6400  
 Max.   :15.90   Max.   :1.5800  
  citric.acid    residual.sugar  
 Min.   :0.000   Min.   : 0.900  
 1st Qu.:0.090   1st Qu.: 1.900  
 Median :0.260   Median : 2.200  
 Mean   :0.271   Mean   : 2.539  
 3rd Qu.:0.420   3rd Qu.: 2.600  
 Max.   :1.000   Max.   :15.500  
   chlorides       free.sulfur.dioxide
 Min.   :0.01200   Min.   : 1.00      
 1st Qu.:0.07000   1st Qu.: 7.00      
 Median :0.07900   Median :14.00      
 Mean   :0.08747   Mean   :15.87      
 3rd Qu.:0.09000   3rd Qu.:21.00      
 Max.   :0.61100   Max.   :72.00      
 total.sulfur.dioxide    density      
 Min.   :  6.00       Min.   :0.9901  
 1st Qu.: 22.00       1st Qu.:0.9956  
 Median : 38.00       Median :0.9968  
 Mean   : 46.47       Mean   :0.9967  
 3rd Qu.: 62.00       3rd Qu.:0.9978  
 Max.   :289.00       Max.   :1.0037  
       pH          sulphates     
 Min.   :2.740   Min.   :0.3300  
 1st Qu.:3.210   1st Qu.:0.5500  
 Median :3.310   Median :0.6200  
 Mean   :3.311   Mean   :0.6581  
 3rd Qu.:3.400   3rd Qu.:0.7300  
 Max.   :4.010   Max.   :2.0000  
    alcohol         quality     
 Min.   : 8.40   Min.   :3.000  
 1st Qu.: 9.50   1st Qu.:5.000  
 Median :10.20   Median :6.000  
 Mean   :10.42   Mean   :5.636  
 3rd Qu.:11.10   3rd Qu.:6.000  
 Max.   :14.90   Max.   :8.000  

summary(white_wine)

 fixed.acidity    volatile.acidity
 Min.   : 3.800   Min.   :0.0800  
 1st Qu.: 6.300   1st Qu.:0.2100  
 Median : 6.800   Median :0.2600  
 Mean   : 6.855   Mean   :0.2782  
 3rd Qu.: 7.300   3rd Qu.:0.3200  
 Max.   :14.200   Max.   :1.1000  
  citric.acid     residual.sugar  
 Min.   :0.0000   Min.   : 0.600  
 1st Qu.:0.2700   1st Qu.: 1.700  
 Median :0.3200   Median : 5.200  
 Mean   :0.3342   Mean   : 6.391  
 3rd Qu.:0.3900   3rd Qu.: 9.900  
 Max.   :1.6600   Max.   :65.800  
   chlorides       free.sulfur.dioxide
 Min.   :0.00900   Min.   :  2.00     
 1st Qu.:0.03600   1st Qu.: 23.00     
 Median :0.04300   Median : 34.00     
 Mean   :0.04577   Mean   : 35.31     
 3rd Qu.:0.05000   3rd Qu.: 46.00     
 Max.   :0.34600   Max.   :289.00     
 total.sulfur.dioxide    density      
 Min.   :  9.0        Min.   :0.9871  
 1st Qu.:108.0        1st Qu.:0.9917  
 Median :134.0        Median :0.9937  
 Mean   :138.4        Mean   :0.9940  
 3rd Qu.:167.0        3rd Qu.:0.9961  
 Max.   :440.0        Max.   :1.0390  
       pH          sulphates     
 Min.   :2.720   Min.   :0.2200  
 1st Qu.:3.090   1st Qu.:0.4100  
 Median :3.180   Median :0.4700  
 Mean   :3.188   Mean   :0.4898  
 3rd Qu.:3.280   3rd Qu.:0.5500  
 Max.   :3.820   Max.   :1.0800  
    alcohol         quality     
 Min.   : 8.00   Min.   :3.000  
 1st Qu.: 9.50   1st Qu.:5.000  
 Median :10.40   Median :6.000  
 Mean   :10.51   Mean   :5.878  
 3rd Qu.:11.40   3rd Qu.:6.000  
 Max.   :14.20   Max.   :9.000  

print("Summary of bank")

[1] "Summary of bank"

summary(bank)

      age            job           
 Min.   :19.00   Length:4521       
 1st Qu.:33.00   Class :character  
 Median :39.00   Mode  :character  
 Mean   :41.17                     
 3rd Qu.:49.00                     
 Max.   :87.00                     
   marital           education        
 Length:4521        Length:4521       
 Class :character   Class :character  
 Mode  :character   Mode  :character  
                                      
                                      
                                      
   default             balance     
 Length:4521        Min.   :-3313  
 Class :character   1st Qu.:   69  
 Mode  :character   Median :  444  
                    Mean   : 1423  
                    3rd Qu.: 1480  
                    Max.   :71188  
   housing              loan          
 Length:4521        Length:4521       
 Class :character   Class :character  
 Mode  :character   Mode  :character  
                                      
                                      
                                      
   contact               day       
 Length:4521        Min.   : 1.00  
 Class :character   1st Qu.: 9.00  
 Mode  :character   Median :16.00  
                    Mean   :15.92  
                    3rd Qu.:21.00  
                    Max.   :31.00  
    month              duration   
 Length:4521        Min.   :   4  
 Class :character   1st Qu.: 104  
 Mode  :character   Median : 185  
                    Mean   : 264  
                    3rd Qu.: 329  
                    Max.   :3025  
    campaign          pdays       
 Min.   : 1.000   Min.   : -1.00  
 1st Qu.: 1.000   1st Qu.: -1.00  
 Median : 2.000   Median : -1.00  
 Mean   : 2.794   Mean   : 39.77  
 3rd Qu.: 3.000   3rd Qu.: -1.00  
 Max.   :50.000   Max.   :871.00  
    previous         poutcome        
 Min.   : 0.0000   Length:4521       
 1st Qu.: 0.0000   Class :character  
 Median : 0.0000   Mode  :character  
 Mean   : 0.5426                     
 3rd Qu.: 0.0000                     
 Max.   :25.0000                     
      y            
 Length:4521       
 Class :character  
 Mode  :character  
                   
                   
                   

The columns job, martial, education, default, housing, loan, contact, month, poutcome, and y need to be converted from characters.

#Bank Dataset

#Fixing Bank Data

#Fixing Martial + Others to Numeric
bank_numeric <- bank %>% mutate(
  marital = case_when(marital == "married" ~ 1,
                      marital == "single" ~ 0,
                      marital == "divorced" ~ -1),
  education = case_when(education == "primary" ~ 1, 
                        education == "secondary" ~ 2, 
                        education == "tertiary" ~ 3, 
                        education == "unknown" ~ NA_real_),
  default = case_when(default == "yes" ~ 1,
                      default == "no" ~ 0),
  housing = case_when(housing == "yes" ~ 1,
                      housing == "no" ~ 0),
  loan = case_when(loan == "yes" ~ 1, 
                   loan == "no" ~ 0),
  contact = case_when(contact == "cellular" ~ 1,
                      contact == "telephone" ~ 2,
                      contact == "unknown" ~ NA_real_),
  poutcome = case_when(poutcome == "success" ~ 1,
                       poutcome == "other" ~ 0,
                       poutcome == "failure" ~ -1, 
                       poutcome == "unknown" ~ 0),
  y = case_when(y == "yes" ~ 1,
                y == "no" ~ 0)
)

#Fixing Months from abb to numbers 

months <- str_to_title(bank$month)
bank_numeric$month <- match(months, month.abb)

#Fixing Job Column 
bank_numeric$job <- as.factor(bank$job)
bank_numeric$job <- unclass(bank_numeric$job)

summary(bank_numeric)

      age             job        
 Min.   :19.00   Min.   : 1.000  
 1st Qu.:33.00   1st Qu.: 2.000  
 Median :39.00   Median : 5.000  
 Mean   :41.17   Mean   : 5.411  
 3rd Qu.:49.00   3rd Qu.: 8.000  
 Max.   :87.00   Max.   :12.000  
                                 
    marital          education    
 Min.   :-1.0000   Min.   :1.000  
 1st Qu.: 0.0000   1st Qu.:2.000  
 Median : 1.0000   Median :2.000  
 Mean   : 0.5019   Mean   :2.155  
 3rd Qu.: 1.0000   3rd Qu.:3.000  
 Max.   : 1.0000   Max.   :3.000  
                   NA's   :187    
    default           balance     
 Min.   :0.00000   Min.   :-3313  
 1st Qu.:0.00000   1st Qu.:   69  
 Median :0.00000   Median :  444  
 Mean   :0.01681   Mean   : 1423  
 3rd Qu.:0.00000   3rd Qu.: 1480  
 Max.   :1.00000   Max.   :71188  
                                  
    housing           loan       
 Min.   :0.000   Min.   :0.0000  
 1st Qu.:0.000   1st Qu.:0.0000  
 Median :1.000   Median :0.0000  
 Mean   :0.566   Mean   :0.1528  
 3rd Qu.:1.000   3rd Qu.:0.0000  
 Max.   :1.000   Max.   :1.0000  
                                 
    contact           day       
 Min.   :1.000   Min.   : 1.00  
 1st Qu.:1.000   1st Qu.: 9.00  
 Median :1.000   Median :16.00  
 Mean   :1.094   Mean   :15.92  
 3rd Qu.:1.000   3rd Qu.:21.00  
 Max.   :2.000   Max.   :31.00  
 NA's   :1324                   
     month           duration   
 Min.   : 1.000   Min.   :   4  
 1st Qu.: 5.000   1st Qu.: 104  
 Median : 6.000   Median : 185  
 Mean   : 6.167   Mean   : 264  
 3rd Qu.: 8.000   3rd Qu.: 329  
 Max.   :12.000   Max.   :3025  
                                
    campaign          pdays       
 Min.   : 1.000   Min.   : -1.00  
 1st Qu.: 1.000   1st Qu.: -1.00  
 Median : 2.000   Median : -1.00  
 Mean   : 2.794   Mean   : 39.77  
 3rd Qu.: 3.000   3rd Qu.: -1.00  
 Max.   :50.000   Max.   :871.00  
                                  
    previous          poutcome       
 Min.   : 0.0000   Min.   :-1.00000  
 1st Qu.: 0.0000   1st Qu.: 0.00000  
 Median : 0.0000   Median : 0.00000  
 Mean   : 0.5426   Mean   :-0.07985  
 3rd Qu.: 0.0000   3rd Qu.: 0.00000  
 Max.   :25.0000   Max.   : 1.00000  
                                     
       y         
 Min.   :0.0000  
 1st Qu.:0.0000  
 Median :0.0000  
 Mean   :0.1152  
 3rd Qu.:0.0000  
 Max.   :1.0000  
                 

hist(bank_numeric$duration, breaks = c(0,10,60,120,300,600,1200,3025))

bank_numeric_lessthan10min <- bank_numeric %>% subset(duration <= 600)
bank_numeric_morethan10min <- bank_numeric %>% subset(duration > 600)
hist(bank_numeric_lessthan10min$duration)

hist(bank_numeric_morethan10min$duration)

boxplot(bank_numeric_lessthan10min$duration)

bank_numeric$y <- as.factor(bank_numeric$y)
x <- bank_numeric %>% subset(select = -c(contact,education,y))
y <- bank_numeric %>% subset(select = c(y))


x_train <- sample_n(x, 0.7*4521)
y_train <- sample_n(y, 0.7*4521)
x_test <- sample_n(x, 0.3*4521)
y_test <- sample_n(y, 0.3*4521)

train <- cbind(x_train, y_train)
test <- cbind(x_test, y_test)

bank_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$y,
                k = 5)
error <- mean(y_test$y != bank_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.87094395280236"

bank_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$y,
                k = 7)
error <- mean(y_test$y != bank_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.881268436578171"

bank_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$y,
                k = 9)
table(y_test$y, bank_knn)

   bank_knn
       0    1
  0 1194    4
  1  158    0

error <- mean(y_test$y != bank_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.880530973451327"

bank_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$y,
                k = 11)
error <- mean(y_test$y != bank_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.883480825958702"

x = rbind(as.matrix(x_train))
y = as.matrix(y_train)

svp <- ksvm(x,y,type="C-svc")
svp_pred <- predict(svp, as.matrix(x_test))
table(svp_pred, as.matrix(y_test))

        
svp_pred    0    1
       0 1198  157
       1    0    1

agreement <- svp_pred == as.matrix(y_test)
prop.table(table(agreement))

agreement
    FALSE      TRUE 
0.1157817 0.8842183 

#plot(svp, data = x)

bank.rf <- randomForest(y~., data = train,
                        ntree = 45,
                        importance = TRUE,
                        proximity = TRUE)
print(bank.rf)

Call:
 randomForest(formula = y ~ ., data = train, ntree = 45, importance = TRUE,      proximity = TRUE) 
               Type of random forest: classification
                     Number of trees: 45
No. of variables tried at each split: 3

        OOB estimate of  error rate: 12.61%
Confusion matrix:
     0  1 class.error
0 2765 12  0.00432121
1  387  0  1.00000000

plot(bank.rf)

pred <- predict(bank.rf, x_test)

table(pred, test$y)

    
pred    0    1
   0 1112  143
   1   86   15

agreement <- pred == test$y
prop.table(table(agreement))

agreement
    FALSE      TRUE 
0.1688791 0.8311209 

importance = importance(bank.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseAccuracy'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip() 

importance = importance(bank.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseGini'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip() 

#Wine Dataset

hist(as.numeric(white_wine$quality))

white_wine$quality <- as.factor(white_wine$quality)
x <- white_wine %>% subset(select = -c(quality))
y <- white_wine %>% subset(select = c(quality))

y <- y %>% mutate(
  quality = case_when(quality == 3 | quality == 4 ~ -1,
                      quality == 5 | quality == 6 | quality == 7 ~ 0, 
                      quality == 8 | quality == 9 ~ 1)
  
)

y$quality <- as.factor(y$quality)
summary(y)

 quality  
 -1: 183  
 0 :4535  
 1 : 180  

x_train <- sample_n(x, 0.7*4898)
y_train <- sample_n(y, 0.7*4898)
x_test <- sample_n(x, 0.3*4898)
y_test <- sample_n(y, 0.3*4898)

train <- cbind(x_train, y_train)
test <- cbind(x_test, y_test)

white_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 5)
error <- mean(y_test$quality != white_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.926480599046971"

white_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 7)
table(y_test$quality, white_wine_knn)

    white_wine_knn
       -1    0    1
  -1    0   59    0
  0     0 1363    0
  1     0   47    0

error <- mean(y_test$quality != white_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.92784206943499"

white_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 9)
error <- mean(y_test$quality != white_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.92784206943499"

white_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 11)
error <- mean(y_test$quality != white_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.92784206943499"

x = rbind(as.matrix(x_train))
y = as.matrix(y_train)

svp <- ksvm(x,y,type="C-svc")
svp_pred <- predict(svp, as.matrix(x_test))
table(svp_pred, as.matrix(y_test))

        
svp_pred   -1    0    1
      -1    0    0    0
      0    59 1363   47
      1     0    0    0

agreement <- svp_pred == as.matrix(y_test)
prop.table(table(agreement))

agreement
     FALSE       TRUE 
0.07215793 0.92784207 

#plot(svp, data = x)

white_wine.rf <- randomForest(quality~., data = train,
                              ntree = 40,
                              importance = TRUE,
                              proximity = TRUE)
print(white_wine.rf)

Call:
 randomForest(formula = quality ~ ., data = train, ntree = 40,      importance = TRUE, proximity = TRUE) 
               Type of random forest: classification
                     Number of trees: 40
No. of variables tried at each split: 3

        OOB estimate of  error rate: 8.52%
Confusion matrix:
   -1    0  1 class.error
-1  0  124  0  1.00000000
0  22 3134 20  0.01322418
1   0  126  2  0.98437500

plot(white_wine.rf)

pred <- predict(white_wine.rf, x_test)

table(pred, test$quality)

    
pred   -1    0    1
  -1    2   28    1
  0    55 1305   46
  1     2   30    0

agreement <- pred == test$quality
prop.table(table(agreement))

agreement
    FALSE      TRUE 
0.1102791 0.8897209 

importance = importance(white_wine.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseAccuracy'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip() 

importance = importance(white_wine.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseGini'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip() 

red_wine$quality <- as.factor(red_wine$quality)
x <- red_wine %>% subset(select = -c(quality))
y <- red_wine %>% subset(select = c(quality))

y <- y %>% mutate(
  quality = case_when(quality == 3 | quality == 4 ~ -1,
                      quality == 5 | quality == 6 | quality == 7 ~ 0, 
                      quality == 8 | quality == 9 ~ 1)
  
)

y$quality <- as.factor(y$quality)
summary(y)

 quality  
 -1:  63  
 0 :1518  
 1 :  18  

x_train <- sample_n(x, 0.7*1599)
y_train <- sample_n(y, 0.7*1599)
x_test <- sample_n(x, 0.3*1599)
y_test <- sample_n(y, 0.3*1599)

red_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 5)
error <- mean(y_test$quality != red_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.937369519832985"

red_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 7)
error <- mean(y_test$quality != red_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.935281837160752"

red_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 9)
table(y_test$quality, red_wine_knn)

    red_wine_knn
      -1   0   1
  -1   0  21   0
  0    0 449   0
  1    0   9   0

error <- mean(y_test$quality != red_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.937369519832985"

red_wine_knn <- knn(train = scale(x_train), 
                test = scale(x_test), 
                cl = y_train$quality,
                k = 11)
error <- mean(y_test$quality != red_wine_knn)
print(paste("Accuracy = ", 1-error))

[1] "Accuracy =  0.937369519832985"

x = rbind(as.matrix(x_train))
y = as.matrix(y_train)

svp <- ksvm(x,y,type="C-svc")
svp_pred <- predict(svp, as.matrix(x_test))
table(svp_pred, as.matrix(y_test))

        
svp_pred  -1   0   1
      -1   0   0   0
      0   21 449   9
      1    0   0   0

agreement <- svp_pred == as.matrix(y_test)
prop.table(table(agreement))

agreement
     FALSE       TRUE 
0.06263048 0.93736952 

#plot(svp, data = x)

train <- cbind(x_train, y_train)
test <- cbind(x_test, y_test)

red_wine.rf <- randomForest(quality~., data = train,
                            ntree = 30,
                            importance = TRUE,
                            proximity = TRUE)
print(red_wine.rf)

Call:
 randomForest(formula = quality ~ ., data = train, ntree = 30,      importance = TRUE, proximity = TRUE) 
               Type of random forest: classification
                     Number of trees: 30
No. of variables tried at each split: 3

        OOB estimate of  error rate: 6.7%
Confusion matrix:
   -1    0 1 class.error
-1  0   49 0  1.00000000
0  12 1044 2  0.01323251
1   0   12 0  1.00000000

plot(red_wine.rf)

pred <- predict(red_wine.rf, x_test)

table(pred, test$quality)

    
pred  -1   0   1
  -1   1  10   0
  0   20 439   9
  1    0   0   0

agreement <- pred == test$quality
prop.table(table(agreement))

agreement
     FALSE       TRUE 
0.08141962 0.91858038 

importance = importance(red_wine.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseAccuracy'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip() 

importance = importance(red_wine.rf)
var_importance = data.frame(Variables = row.names(importance),
                           Importance =round(importance[, 'MeanDecreaseGini'],2))
rank_importance=var_importance %>%
  mutate(Rank=paste('#',dense_rank(desc(Importance))))
ggplot(rank_importance,aes(x=reorder(Variables,Importance),
 y=Importance,fill=Importance))+ 
 geom_bar(stat='identity') + 
 geom_text(aes(x = Variables, y = 0.5, label = Rank),
 hjust=0, vjust=0.55, size = 4, colour = 'white') +
 labs(x = 'Variables') +
 coord_flip()