Travel Insurance Recommendation AI System

This repository contains the implementation of an AI system designed to predict flight delays and recommend travel insurance to customers based on their potential purchase intentions.

Overview

Our AI system, Insurance-GPT, leverages a Large Language Model (LLM) to analyze customer interactions in real-time and predict flight delays using a deep learning model. This system provides personalized insurance recommendations to improve user experience and offers valuable insights for insurance pricing strategies.

Features

• Real-time Interaction: Insurance-GPT interacts with users to understand their needs and preferences.
• Flight Delay Prediction: Utilizes a deep learning model to predict flight delays accurately.
• Personalized Insurance Recommendations: Provides tailored insurance suggestions based on predicted delays and customer sentiment.

Datasets

We used three main datasets to train and evaluate our models:

1. Travel Insurance Dataset: Contains customer profiles and their potential to purchase insurance.
2. Twitter US Airline Sentiment Dataset: Provides user reviews about US airlines, categorized as positive or negative.
3. InsuranceQA-v2 Dataset: Contains insurance-related questions and answers for training our model.

Model Implementation

LLM Fine-Tuning

1. Supervised Fine-Tuning (SFT): On the insuranceQA-v2 dataset to improve domain-specific understanding.
2. Odds Ratio Preference Optimization (ORPO): To enhance the model's generalization and alignment with human preferences.
3. LoRA Fine-Tuning: On the Travel Insurance Dataset to improve precision in predicting user purchase intent.

Deep Learning Model

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.

Citation

If you use this code or dataset in your research, please cite our report using the following BibTeX entry:

@techreport{yang2024insurancegpt,
  author       = {Yuzhe Yang and Haoqi Zhang and Zhidong Peng and Yilin Guo and Tianji Zhou},
  title        = {Travel Insurance Recommendation AI System Based on Flight Delay Predictions and Customer Sentiment},
  institution  = {The Chinese University of Hong Kong, Shenzhen},
  year         = {2024},
  url          = {https://github.com/TobyYang7/Travel-Insurance-Recommendation-AI-System},
  note         = {Available at: \url{https://github.com/TobyYang7/Travel-Insurance-Recommendation-AI-System}}
}