Project information
- Category: Sentiment Analysis & Classification
- Client: Georgia State University(Research Project)
- Project date: 05 Dec, 2022
Project Description
- Main aim of this research project is to classify extremist and non-extremist tweets
- Extracted 100K Tweets using Twitter API, performed data cleaning to remove special characters, links and loaded the data to PostgresDB
- Created N-grams and visualized top 5 topics using LDA model. Created word embeddings using Word2Vec and used T-SNE for visualzing the embeddings
- Performed finetuning of the word embeddings to classify extremist and non extremist tweets using BERT and GPT-2