CUHKSZ 24-Spring IBA4315 Final Project, ASTGCN+LLM
This project will design a new Al system to predict flight delays and price the insurance, enabling the airlines can provide personalized insurance recommendations for passengers. This Al system will analyze the rates of flight delays based on multiple relevant data, such as flight dynamic data, city weather, and special situations data. Besides, it will construct user profiles with sentiment analysis based on users' feedback to airlines. Finally, the system will offer a personalized and optimal insurance recommendations, contributing to supplying a suitable service for each customer and gaining more benefits for airlines. Through personalized insurance recommendations, our AI system can boost customer satisfaction and loyalty while reducing stress caused by flight delays. Our system could reshape customer service in the airline industry and promote a positive image and competitive edge for airlines, resulting from focusing on prediction accuracy and user interaction quality.
Yuzhe Yang, Haoqi Zhang, Zhidong Peng, Yilin Guo, Tianji Zhou
United States Department of Transportation
Our deep learning model uses the ASTGCN framework to predict flight delays based on spatio-temporal data. This framework captures both temporal and spatial correlations in the data for accurate predictions.
Travel Insurance Recommendation AI Systerm Based on Flight Day Predictions and Customer Sentiment