|
The AI-Powered Virtual Assistant Chatbot is aimed at improving the accessibility and efficiency of public sector engagement by delivering automated answers to citizens' questions regarding government schemes and services. Based on Natural Language Processing (NLP) and Machine Learning (ML), the chatbot smartly analyzes user queries, categorizes them into pre-defined buckets, and fetches correct information from a structured JSON database. The framework uses a multi-step process, focusing on tokenization, classification based on ML models (Neural Networks, SVM, Decision Trees, and Random Forest), response retrieval, and presentation.Performance tests show excellent classification accuracy, with Neural Networks performing at 92% accuracy, surpassing other models. The chatbot shows an average response time of 2.2 to 3.1 seconds, guaranteeing fast and consistent information dissemination. In spite of its performance, issues like dealing with vague queries persist, as seen in increased error rates for categories such as "Benefits and Advantages" (9.0%) and "Status Tracking" (8.5%).Future enhancements will aim to maximize query classification accuracy and response generation speed to further improve user satisfaction. The system proposed has the potential to alleviate the burden on government service centers, providing efficient, timely, and user-friendly access to vital information for citizens.
Keywords: AI Chatbot, NLP, ML, Government Schemes, Citizen Services, Query Classification, Tokenization and Neural Networks
|
