Interpretable noninvasive diagnosis of tuberculous pleural effusion using LGBM and SHAP: development and clinical application of a machine learning model

View article
PeerJ

Main article text

 

Introduction

Materials and Methods

Data collection protocol

Data sources

Data application and processing

Feature selection

Machine learning model construction

Machine learning algorithms

Model performance evaluation

Assessment metrics for a diagnostic model

Statistical analysis

Results

Baseline analysis

Model development and performance comparison

LGBM model performance by feature set

Final model identification

External validation and performance evaluation of feature parameters

Model interpretability

Practical clinical application

Discussion

Sample size and multicenter validation

Multimodal data integration

Limitations and future outlook

Future research directions

Cost-effectiveness analysis

Conclusion

Supplemental Information

Confusion Matrix of Internal Validation.

DOI: 10.7717/peerj.19411/supp-2

Confusion Matrix of External Validation.

DOI: 10.7717/peerj.19411/supp-3

Additional Information and Declarations

Competing Interests

Author Contributions

Human Ethics

Data Availability

Funding

This work was supported by the Scientific Research Fund of Zhejiang Provincial Education Department, China (No. Y202045308); the fund of the Zhejiang Province Medical and Health Science and Technology Project, China (No. 2022KY877); the fund of the Jiashan Bureau of Science and Technology, China (No. 2023A60). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.