원문정보
Predicting Travel Insurance Adoption among Chinese Travelers : A Machine Learning and Deep Learning Approach
초록
영어
Despite the rapid growth of outbound tourism in China, the adoption of travel insurance remains limited. This study investigates the determinants of insurance purchase among 2,000 Chinese travelers using machine learning and deep learning techniques. The dataset includes demographic, socioeconomic, and behavioral variables such as age, income, family size, health status, and travel experience. Four models—logistic regression, Random Forest, a baseline deep neural network (DNN), and a dropout- enhanced DNN—were developed and compared. Results show that the dropout-enhanced DNN achieved the highest accuracy (about 82%), outperforming the baseline DNN (79%), Random Forest (79%), and logistic regression (75%). Feature importance analysis indicated that annual income, age, and family size are the most decisive predictors of adoption. The findings highlight the value of artificial intelligence in predicting consumer behavior in the insurance sector. For practitioners, the results suggest that insurers should target higher-income and family-oriented travelers while leveraging airlines and agencies as key distribution channels. Predictive analytics can thus support more effective segmentation, targeting, and personalized insurance design.
목차
1. Introduction
2. Related Research
2.1 Travel Insurance and Consumer Behavior
2.2 Machine Learning and Deep Learning Applications
3. Research Methodology
3.1 Data Collection and Preprocessing
3.2 Classification Model Development
3.3 Evaluation Strategy
4. Experimental Results
4.1 Classification Model Performance
4.2 Key Predictors of Insurance Adoption
4.3 Summary of Findings
5. Conclusion and Business Implications
5.1 Conclusion
5.2 Business Implications
References