[Application of principal component analysis in comprehensive indicator screening for pneumoconiosis in different...

Objective: To investigate the application of principal component analysis in comprehensive indicator screening for pneumoconiosis in different regions. Methods: A principal component analysis was performed for the data of 11 factors associated with the prediction of pneumoconiosis hazard and collected in the investigation on occupational health status conducted in 172 counties (districts) in Hebei, China. The degree of pneumoconiosis hazard in different regions was obtained and intuitively presented by GIS. Results: The eigenvalues of 5 principal components of pneumoconiosis were 4.103, 2.341, 0.981, 0.943, and 0.726, respectively, and the contribution values were 37.299%, 21.286%, 8.919%, 8.572%, and 6.596%, respectively. According to the comprehensive value of principal components, GIS Natural Breaks was used to divide the degree of pneumoconiosis hazard in 172 counties (districts) in Hebei into mild, moderate, and severe grades. Of all the counties, 46 had severe pneumoconiosis hazard, 69 had moderate pneumoconiosis hazard, and 57 had mild pneumoconiosis hazard, and the ranges of the score of principal components were 0.30-1.15, -0.24 to 0.27, and -0.69 to -0.25, respectively. Conclusion: Principal component analysis can optimize the comprehensive indicators for the evaluation of regional pneumoconiosis. The comprehensive score of principal components can quantify and intuitively show the degree of pneumoconiosis hazard in different regions. Tangshan, Chengde, Shijiazhuang, and Handan have the most severe pneumoconiosis hazard.