Intelligent Chatbot for Educational Communication and Support

This study explores the development of an intelligent response system that leverages machine learning models to generate context-specific answers from a knowledge base. The system aims to enhance decision-making and information retrieval in a university setting, where quick and accurate responses to queries are vital. The framework consists of three key components: Raw Data Input, Train Data, and Response. Raw Data Input represents the independent variable, which undergoes several preprocessing steps—keyword spotting, verification, tokenization, and semantic extraction—before being passed to the Train Data stage. The Train Data serves as a moderator, where the clean data is processed using two models: the Sequence to Sequence (seq2seq) Model and the Information Retrieval (IR) Model. The seq2seq Model is designed to generate responses that are outside the scope of the school's knowledge base, while the IR Model generates answers based on the existing knowledge base, ensuring domain-specific relevance. The study focuses on analyzing the relationship between these variables, emphasizing how the input data and models influence the system's ability to provide accurate, context-aware responses. The results indicate that the integrated approach, combining Keyword and Intent Spotting to the seq2seq and IR models, significantly improves the response accuracy and contextual relevance compared to traditional methods. This research contributes to the field of intelligent systems, particularly in education, by providing insights into effective data processing and response generation techniques. The paper concludes with recommendations for future enhancements and applications of the proposed system.