Skip to content

Latest commit

 

History

History
7 lines (5 loc) · 404 Bytes

README.md

File metadata and controls

7 lines (5 loc) · 404 Bytes

Bengali_news_classifier

Classifying Bengali news articles based on their headlines using NLP

Results

News headlines from anadabazar, zeenews and ebala was used to train a Bengali NLP model on 5 classes: Kolkata, State, National, Entertainment and Sport.

The model acheived a training accuracy of 92.98% and a testing accuracy of 90.19%. A training F1 score of 0.81 and testing of 0.76 was acheived