Data set construction method and device for prevention and treatment of occupational hearing loss

By using data association matching, feature filtering, and multi-factor analysis, a high-quality dataset of occupational hearing loss was constructed, which solved the problems of low efficiency in the fusion of multi-source heterogeneous data and inconsistent data quality, and achieved high-precision prediction of early risks and verification of model credibility.

CN122177500APending Publication Date: 2026-06-09EAST CHINA UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
EAST CHINA UNIV OF SCI & TECH
Filing Date
2026-05-09
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In existing technologies, the sources of occupational hearing loss are wide and heterogeneous, resulting in low data fusion efficiency and inconsistent quality. There is a lack of high-precision early risk prediction models, and existing models have limited generalization ability and prediction accuracy when dealing with nonlinear coupling relationships among multiple factors.

Method used

By matching data from different sources and combining professional theories of occupational health medicine and hearing loss prevention, feature screening and preprocessing were performed to identify key feature subsets. Multifactor correlation analysis and oversampling balance were used to train multiple machine learning models, select the optimal model, and verify its consistency with the medical theory of noise-induced indentation.

Benefits of technology

A high-quality standard dataset was constructed, enabling early quantitative prediction of hearing loss risk, improving prediction accuracy and model credibility, and meeting compliance requirements in the field of occupational health.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122177500A_ABST
    Figure CN122177500A_ABST
Patent Text Reader

Abstract

This application provides a method and apparatus for constructing a dataset for the prevention and treatment of occupational hearing loss. The method includes obtaining fused data by data association and matching from raw data from different sources, and incorporating medical theory into various stages such as feature selection and data preprocessing of the fused data. This makes the preprocessing results more targeted, supporting the task of preventing and treating occupational hearing loss. By performing multi-factor correlation analysis on a standard dataset, a subset of key features most strongly associated with hearing loss is selected. The standard dataset is oversampled and balanced to obtain a balanced dataset, which is then used to construct and train multiple machine learning models to select the optimal model. The high-precision performance of the optimal model on the test set verifies that the standard dataset has the ability to support highly robust predictions. Furthermore, using a suitable feature importance quantification method, the optimal model is verified to be consistent with medical theoretical expectations, thereby enhancing the credibility and interpretability of the prediction results.
Need to check novelty before this filing date? Find Prior Art