Naive Bayes

Question 1

What is it for?

Answer 1

Classification only.

It’s used when all features have equal weight, and are independent.

Question 2

What do we have to do if predicted values are 0 and 1?

Answer 2

We have to tell the classifier that 0 and 1 are categorical data. we use as.factor()

Question 3

How to find accuracy in R?

Answer 3

We take the mean of prediction compared to actual values:

mean(prediction == testset$)

Question 4

Is Data Scaling necessary?

Answer 4

There is no need to scale the data in Naive Bayes.

Question 5

How do we predict new inputs in R?

Answer 5

• The input needs to be Class data.frame.
• then, we use the function predict, with the Classifier and the input data (minus the output).

nb = NaiveBayes(target ~ ., data=trainset)
predict(nb, new_input[:, -col_num]) # col_num is result