New energy vehicle insurance premium prediction method and device based on cross-domain transfer learning, computer equipment and storage medium

Through cross-domain transfer learning methods, combined with feature splicing and adversarial network training of new energy vehicles and fuel vehicles, a new energy vehicle premium prediction model was constructed, which solved the problem of inaccurate risk assessment of electric vehicles and improved the accuracy and stability of premium prediction.

CN120746731APending Publication Date: 2025-10-03SHANGHAI QIHAO INFORMATION TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202510847688.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-24
Publication Date
2025-10-03

AI Technical Summary

Technical Problem

Existing technologies are unable to accurately assess the risk status of electric vehicles, resulting in inaccurate predictions of new energy vehicle premiums. The lack of electric vehicle-specific data and related driving behavior data makes it challenging to build an accurate electric vehicle risk prediction model.

Method used

A cross-domain transfer learning method is adopted to obtain new energy vehicle data, extract feature vectors, and splice them with the time series feature vectors of fuel vehicles. The underlying parameter freezing and fine-tuning of the feature fusion layer and premium prediction layer of the fuel vehicle pre-training model are utilized to construct a new energy vehicle premium prediction model. The adversarial network and gradient reversal layer are combined for training to reduce the difference in feature distribution.

Benefits of technology

It improves the accuracy and stability of new energy vehicle premium prediction, enhances the generalization ability of the model, can better capture the risk characteristics of new energy vehicles, and realizes cross-domain knowledge transfer.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the field of vehicle insurance risk prediction, and discloses a new energy vehicle insurance premium prediction method and device based on cross-domain transfer learning, computer equipment and a storage medium, and the method comprises the steps: obtaining new energy vehicle data; extracting features of the new energy vehicle data to obtain new energy vehicle feature vectors; determining a splicing feature vector according to the new energy vehicle feature vector and the fuel vehicle time sequence feature vector; and processing the spliced feature vector through a new energy vehicle premium prediction model based on cross-domain transfer learning, and obtaining premium prediction data output by the new energy vehicle premium prediction model. According to the method, the new energy vehicle characteristics and the fuel vehicle time sequence characteristics are fused, and the accuracy of insurance premium prediction is improved by using cross-domain transfer learning; the problems that new energy vehicle data is relatively insufficient and independent modeling is difficult are effectively solved; through feature splicing and transfer learning, the new energy vehicle insurance premium prediction model can better capture the risk features of the new energy vehicle by referring to the historical law of the fuel vehicle.
Need to check novelty before this filing date? Find Prior Art