Welcome to the User Generated Bug Reports using Topic Modelling repository! This project aims to automatically identify topics in user-generated bug reports and reviews for software applications using topic modelling techniques. By extracting key themes and issues from user feedback, developers can prioritize and improve their software based on user needs and preferences.
This project utilizes BERTopic, a state-of-the-art topic modelling algorithm based on BERT (Bidirectional Encoder Representations from Transformers), to automatically identify topics in user-generated bug reports and reviews for software applications.
BERTopic uses advanced natural language processing techniques to preprocess and clean raw text data, and then applies the BERT language model to generate embeddings, which are then clustered to form coherent topics. The output of this model is a list of topics and their associated reviews/documents, which can be used to identify and categorize user feedback.
In addition to topic modelling, we also employ sentimental analysis to understand how users feel about bug and feature related issues. This information can be used to further prioritize and address user concerns.