2

dataset can be find here : https://archive.ics.uci.edu/ml/datasets/Bank+Marketing#

set.seed(1234)
ind <- sample(2, nrow(bank), replace = TRUE, prob = c(0.7, 0.3))
train.data <- bank[ind == 1, ]
test.data <- bank[ind == 2, ]

I was searching my problem i tried to set it to factor in confusion matrix. But problem didn't solved at all

cartmodel <- rpart(y ~., data = train.data)
cartmodel
cart.pred = predict(cartmodel, test.data)
summary(cart.pred)
confusionMatrix(as.factor(cart.pred),as.factor(test.data$y))
confusionMatrix

What i need to change? Dataset is Bank.. so Num and Factor attributes.

Update : i tried change all atributed to factor.. still error

StupidWolf
  • 45,075
  • 17
  • 40
  • 72
Dzurko Music
  • 25
  • 5
  • Please provide an example of your data set in order to make it reproducible. If the data set is publicly available, please include a description on how to access it. – Phil Jan 31 '21 at 17:54
  • Dataset is from UCI .. https://archive.ics.uci.edu/ml/datasets/Bank+Marketing# – Dzurko Music Jan 31 '21 at 18:04

1 Answers1

1

Using the csv from UCI (can also try this link):

library(rpart)
library(caret)
bank = read.csv("../bank/bank-full.csv",sep=";")

set.seed(1234)
ind <- sample(2, nrow(bank), replace = TRUE, prob = c(0.7, 0.3))
train.data <- bank[ind == 1, ]
test.data <- bank[ind == 2, ]

When you call predict() you are getting the probabilities, not the labels:

cartmodel <- rpart(y ~., data = train.data)
cart.pred = predict(cartmodel, test.data)
head(cart.pred)

          no        yes
5  0.9393461 0.06065387
14 0.9393461 0.06065387
16 0.9393461 0.06065387
26 0.9393461 0.06065387
28 0.9393461 0.06065387
29 0.9393461 0.06065387

To get labels:

cart.pred = predict(cartmodel, test.data,type="class")
confusionMatrix(cart.pred,test.data$y)

Confusion Matrix and Statistics

          Reference
Prediction    no   yes
       no  11710  1039
       yes   302   517
                                         
               Accuracy : 0.9012         
                 95% CI : (0.896, 0.9061)
    No Information Rate : 0.8853         
    P-Value [Acc > NIR] : 1.831e-09      
                                         
                  Kappa : 0.3869
StupidWolf
  • 45,075
  • 17
  • 40
  • 72