Legal Query Retrieval System for Indian Law Students

Legal researchers search large legal texts to find the most relevant information. Searching for the desired legal information by hand is inefficient and time-consuming. For this reason, this research proposes an AI-based legal input retrieval system that allows users to retrieve appropriate legal provisions from the Narcotic Drugs and Psychotropic Substances (NDPS) Act of 1985. The system utilises machine learning and natural language processing techniques to assist users in retrieving the correct legal sections from the NDPS Act based on user input. The system implements query classification using TF-IDF vectorization and Logistic Regression and utilises Sentence-BERT based embeddings and cosine similarity for semantic retrieval. The implementation of the system is in Python and is deployed with a Streamlit web application interface. The experimental results indicate that the AI-based legal retrieval system improved accuracy of retrieval and reduced the time taken for retrieving legal information in comparison to a traditional