With the rapid development of e-commerce, the global e-commerce market has shown a thriving trend, with the scale exceeding $6.5 trillion in 2024. However, behind this growth, customer complaints and churn issues have become increasingly prominent. Data shows that online shopping complaints accounted for 22.3% of total complaints in 2024, and improper handling of complaints can lead to customer churn rates as high as 18%. Customer churn not only reduces revenue but also damages brand image and market reputation, making it a key factor restricting the long-term development of e-commerce enterprises. 

Against this backdrop, how to accurately predict customer complaints and churn risks, and take scientific intervention measures has become an urgent problem for e-commerce platforms. Traditional prediction models often face limitations such as over-reliance on single algorithms, insufficient utilization of unstructured data, and poor adaptability to complex data distributions. These defects make it difficult to meet the actual needs of enterprises for precise customer management. 

Download Full Text