Nepali News Classifier is an application that can classify the unlabelled or unclassified news/text document into predefined target categories. Following steps were followed in making this project:
- First Nepali News Datasets were prepared by extracting News articles from various Nepali News Portals usingn BeautifulSoup library of Python
- 12 Classes of News were predefined (Sports, Health, World, Politics etc.) and news extraction was done in each category for train-test purpose
- Scikit-learn library was used for training and model building, training news datasets were trained with different supervised learning algorithms available in scikitlearn (svm,logistic regression, naive bayes etc)
- Model Performances were evaluated with test datasets (82% accuracy was the best) 5.Best Model were used in the final application (GUI: tkinter and Web: Flask framework)