I'm using logistic regression these days.
In order to construct ROC curve, I need decision values of logistic
How can I get decision values ( data * LR_coefficients ) of logistic
regression through weka.
Any help will be appreciated.
Thanks very much.
I,m from china.
Recently,I use weka to train model,and use it to classify.
The source file format is like 'file.txt' in attachment,and with arff/model file.
I run the cmd:
java weka.classifiers.bayes.NaiveBayes -t file.arff
the result is:
=== Stratified cross-validation ===
Correctly Classified Instances 6300 78.583 %
Incorrectly Classified Instances 1717 21.417 %
Kappa statistic 0.1092
Mean absolute error 0.2161
Root mean squared error 0.4486
Relative absolute error 195.5779 %
Root relative squared error 190.9402 %
Coverage of cases (0.95 level) 82.6494 %
Mean rel. region size (0.95 level) 54.4281 %
Total Number of Instances 8017
=== Confusion Matrix ===
a b <-- classified as
6098 1449 | a = 0
268 202 | b = 1
It means success!
The question is how to classify single file (txt format ,such as 'file.txt') use weka,what's the cmd?
Please help me,Thx!