how to calculate model accuracy in rstudio for logistic regression

user_15192336 0
Viewed 533

How do you calculate the model accuracy in RStudio for logistic regression. The dataset is from Kaggle.

set.seed(1000)
split = sample.split(query$Exited, SplitRatio = 0.65)
train = subset(query, split==TRUE)
test = subset(query, split==FALSE)
model = glm(Exited ~ CreditScore + Gender + Age + Balance + IsActiveMember, data = train, family=binomial)
summary(model)
predict = predict(model, type="response", newdata=test)
table(test$Exited, predict > 0.5)

FALSE TRUE

0 2717 70

1 606 107

Is it possible to extract the values from the table to calculate the accuracy using variables or is there a function to get the accuracy?

# Accuracy of model:
(2717+107)/(2717+70+606+107)

Is it more accurate? I'm getting different values.

(2717+107)/(2717+70+606+107)

accuracy is 0.8068571

Accuracy(y_pred = pred, y_true = train$Exited)

accuracy is 0.8087692 using ML metrics

logistic-regression r
1 Answers

use the package ML metrics

MLmetrics::Accuracy(predicted, actual)

There is also a lot of different error methods you can calculate within that package

user_120439460
Related