This was created as a final year project for CsIT.
Multiple news websites were scraped to collect the headlines. We labeled them as Negative and Positive. News headlines with neutral sentiment were also labeled Positive.
complete.ipynb is the Jupyter notebook file where the complete code is kept.
Overall accuracy, recall and precision was around 82%.
api.py contains the Flask code where the model was implemented to scrape Kantipur Daily and other sites in real time to classify the sentiment of the headlines.
Members:
Ayush Pandey, Arati Bhattarai, Kushal Lal Shrestha, Dip Aryal
NCCS College, Year 2019