CODEWITHSUNDEEP
classificationwekanaivebayesrocmlp
Encipher
Encipher

Reputation: 2936

How to create ROC curve for different classifier in Weka or excel

I have a array of sensitivity and specificity values for positive class for different classifiers. I want to create one ROC curve for each classifier.

For example

                   Sensitivity    specificity            ROC
 NB                 0.613         0.778                  0.791
 LR                 0.865         0.842                  0.88
 MLP                0.976         0.903                  0.959

Those are not real value I created those value for demonstration purpose. I mention here sensitivity and specificity because ROC is the ratio of True positive and False positive rate.

I want a plot like that

enter image description here

I also go through the Weka Tutorial 30: Multiple ROC Curves (Model Evaluation). The knowledge flow diagram he was talking about had two drawback

1. If I have a training and test dataset and I want to see the ROC of test dataset that was not defined there.

2. If I am using 5 fold cross validation on training set how could I represent that, that was also not defined.

I tried to make a knowledge flow environment by myself but I did not get the option on arff loader "the load model".

Upvotes: 1

Views: 141

Answers (0)

Related Questions