Model training method and device, computer device, and storage medium

By filtering and retraining the training data, and combining the priority of multiple models, the problem of insufficient accuracy in the training of existing models is solved, and more efficient model training and recognition accuracy are achieved.

CN115345299BActive Publication Date: 2026-05-29CHINA PING AN LIFE INSURANCE CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
CHINA PING AN LIFE INSURANCE CO LTD
Filing Date
2022-08-22
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing model training methods suffer from insufficient model accuracy and low training efficiency, making it difficult to effectively improve through algorithm improvements or model fusion.

Method used

By dividing the training data into a training set and a test set, N general models are trained using the training set, and data with significant features are selected from the test set for secondary training to obtain N secondary classification models. The model priority is determined based on the model test results, and the models are combined into a comprehensive classification model.

Benefits of technology

This improved the model's accuracy and training efficiency, resulting in a more accurate comprehensive classification model and enhancing the accuracy of the model's recognition results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115345299B_ABST
    Figure CN115345299B_ABST
Patent Text Reader

Abstract

The application relates to the technical field of artificial intelligence, and discloses a model training method and device, computer equipment and a storage medium. The method comprises the following steps: first, N general models are trained by using first training data in a training set, and N first classification models are obtained; second, the N first classification models are tested by using the first training data in a test set, so as to obtain model test results corresponding to the N first classification models, the model test results are used for screening the first training data in the test set, and second training data with more obvious characteristics is obtained; finally, the N first classification models are trained by using the second training data, N second classification models are obtained, and the N second classification models are combined to obtain a comprehensive classification model with higher precision, so that the problem of insufficient precision of a model obtained by model training is solved.
Need to check novelty before this filing date? Find Prior Art