CODEWITHSUNDEEP
rlogistic-regressionconfusion-matrix
user3230594
user3230594

Reputation:

Logistic regression confusion matrix

here is part of my code below for an assignment I am working on this semester:

fit2=glm(card~reports+income+age+owner+dependents+months+share, data=new_credit2, family="binomial")
summary(fit2)


####Part G####

pred_prob=predict(fit2,type="response")
head(pred_prob)

length(pred_prob)

# The contrasts() function indicates that R has created a dummy variable with a 1 for =Yes

contrasts(card)


# The following command creates a vector of 1,319 No elements

glm.pred=rep("No",1319)


#The following command transforms all the elements with predicted probabilities of acceptance 
greater than 0.5 from No to Yes

glm.pred[pred_prob>.5]="Yes"

head(glm.pred)

head(card)

#table() produces a confusion matrix to determine how many observations were correctly or 
incorrectly classified

table(glm.pred,card)


# mean(): computes fraction of individual for which the prediction was correct
mean(glm.pred==card)

When I run this I get a matrix that looks like this:

         card
glm.pred  no yes
     No   86 232
     Yes 210 791

Yet, when I run the mean() function to try to get the fraction of correct predictions, I get a result of 0. I am unsure why this is occurring, and was hoping someone could lead me in the right direction.

Thanks everyone

Upvotes: 0

Views: 330

Answers (1)

mnist
mnist

Reputation: 6954

If this is really your output, notice the different spelling of Yes - yes and No - no. Cheers

Upvotes: 1

Related Questions