Real Estate Chatbot Assistant: Simplifying Property Guidance for Beginners

Abstract

This study presents the development of a chatbot designed to assist users with real estate consultations, focusing on comparing two advanced language models, Llama 3.1 and OpenThaiGPT, to identify the most suitable one for this purpose. A dataset of 50 real estate-related documents, including topics such as mortgages, legal regulations, and market trends, was curated and processed to train the chatbot. MPNet embeddings were used to ensure the chatbot could understand and respond accurately to user queries, while a reranking system prioritized the most relevant results. LangChain was integrated to improve the system’s ability to provide clear and structured responses. The chatbot was evaluated using user feedback and performance metrics such as BLEU and ROUGE to measure its accuracy and fluency. The findings demonstrate that the chatbot is highly effective in assisting users with limited knowledge of real estate by providing straightforward and accessible advice. Its ability to clarify processes, explain key concepts, and guide users through real estate decisions makes it a valuable tool for those seeking guidance in buying, selling, or understanding the market.

Ähnliche Arbeiten