Sentiment-Analysis

Deep Learning model for Sentiment Analysis trained using steam reviews and designed to predict on Twitter data.
Used GloVe embedding layer and lstms to create a Neural Network that can classify tweets into a range between negative(0) and positive(1).
The model was trained on Steam Reviews of 3 games: Witcher 3, Cyberpunk, Gwent and then was used to predict on tweets around the time of release of these games.
Model acheived a validation accuaracy of 85% and report and analysis was prepared based on the sentiment predictions of tweets.
Report
Presentation