AI FAQ Chatbot (training on labelled query-response pairs)
Built and trained a rule-based NLP FAQ chatbot using a dataset of 150+ labelled query-response pairs. Implemented tokenisation and response matching using NLTK to handle diverse student questions. Evaluated chatbot performance by measuring accuracy across 50 test queries. • Created labeled training pairs in a question-to-response format • Preprocessed and tokenized queries for model inference • Performed response selection for new, unseen questions • Conducted accuracy testing across multiple query types