Customer emotion feature extraction method and system based on service data

By adopting a learning framework and dynamic adjustment strategy with joint supervision of multiple teacher models in customer emotional feature extraction, the problem of insufficient labeling and noise is solved, the extraction accuracy and model generalization ability is improved, applicable to different business scenarios and improved robustness .

CN119202649AInactive Publication Date: 2024-12-27BEIJING HUADANG ZHIYUAN TECH CO LTD
View PDF 4 Cites 1 Cited by

Patent Information

Application Number
CN202410500198.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2024-03-26
Filing Date
2024-04-24
Publication Date
2024-12-27
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure CN119202649A_ABST
    Figure CN119202649A_ABST
Patent Text Reader

Abstract

The invention discloses a customer emotion feature extraction method and system based on service data, and the method comprises the steps: obtaining first service data, carrying out the data cleaning of the first service data, and defining the first service data after the data cleaning as a data training set; a common supervised learning framework is constructed, the common supervised learning framework comprises a plurality of teacher models and a student model, and the plurality of teacher models jointly supervise the student model; training the student model based on the data training set and the common supervised learning framework, the student model being supervised by different teacher models in the training process; and obtaining second service data, and based on the second service data, performing customer emotion feature extraction through the trained student model.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Multi-modal sentiment analysis method based on multi-task learning and attention layer fusion

    CN114330551A

  • Customer service monitoring method and device based on voice and expression emotion recognition

    CN115862101A

  • Multi-teacher knowledge distillation method based on reinforcement learning

    CN117391180A

  • Knowledge distillation method and system based on multi-teacher multi-modal model

    CN117669693A

Cited By

  • Classroom event analysis method and analysis model optimization method and system

    CN121415298A