In this project we worked in a team of five people. This was assigned to us in the course- Data Science and Analytics. This project was also a part of national level data science competition in Canada.
During the course of project we developed three classification models to predict the truth ratings (labels) that human fact checkers would assign to each claim based on some related articles and the metadata. The models include Passive Aggressive algorithm, LSTM and BERT. The comparison for models with different features on the basis of F1-Score and Train, Test Accuracy was done towards the end to select the model with best performance