Projekt

AI-Based Student Advisory Chatbot

This project focuses on developing an AI-based chatbot to support students with questions about their study programs.
The university’s current chatbot is limited, as it mainly relies on predefined rules and keyword-based matching and responds in a very rigid and unsatisfactory way.
The goal of this project is to enable a chatbot that not only understands student questions more naturally, but also provides answers based on official documents such as module handbooks and study regulations.
It is designed as an extension of the university’s existing chatbot, aiming to make future query answers more accurate, flexible, and context-aware.
To achieve this, the project uses a Retrieval-Augmented Generation (RAG) approach combined with modern NLP techniques.
PDF documents are processed by extracting text and identifying relevant information such as modules and credits, which is then stored as structured JSON data.
The data is further prepared for semantic search by splitting it into smaller text sections and creating embeddings.
This allows the chatbot to understand the meaning of user queries and retrieve relevant information even when questions are phrased differently.
The retrieved information is then used together with a language model to generate context-aware responses.
In addition, the system provides transparency by displaying the sources used to generate each answer.