Text Mining

I have a problem in Text Classification.

After classifying a Sample news manually into FIVE Groups like

Highly Positive , Positive , Neutral , Negative & Highly Negative number of Samples are small in Highly Positive & Highly Negative category.

Almost all classifiers including XgBoost , Random Forest are misidentifying the Samples with less representation.

How to get over this problem ?

This sounds like the classic rare event problem. You might check the threads I’ve linked below - this is something that comes up now and again: