Training And testing the SVM Classifier for sentiment analysis of Text. The input file contains CSV file containing 2 columns : Text and Sentiment (POS OR NEG). The process works fine with Decision Tree and Naive Bayes but gives "SVM Learner rejecting 1 column due to incompatible type" error in SVM Learner. I am attaching the screenshot of my process here.
this simly means that the SVM Learner can not handle the "Document" column and ignores it during training, which is fine. The SVM Learner should use the numerical columns as features and the sentimen as target. The Document column is not of any use anymore, since the model is build based on the vector representation.
