Regression analysis for solving diagnosis problem of children's health

International Conference on Mechanical Engineering, Automation and Control Systems 2015

机械制造;无线电电子学;计算机科学

Cherkashina, Yu.A.^1 ; Gerget, O.M.^1

Tomsk Polytechnic University, 30, Lenina ave., Tomsk

634050, Russia^1

关键词: Binary logistic regression;    Evidence-based medicine;    Practical importance;    Professional activities;    Scientific researches;    Sensitivity and specificity;    Statistical techniques;    Vascular endothelial growth factor;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/124/1/012047/pdf
DOI  :  10.1088/1757-899X/124/1/012047

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

The paper includes results of scientific researches. These researches are devoted to the application of statistical techniques, namely, regression analysis, to assess the health status of children in the neonatal period based on medical data (hemostatic parameters, parameters of blood tests, the gestational age, vascular-endothelial growth factor) measured at 3-5 days of children's life. In this paper a detailed description of the studied medical data is given. A binary logistic regression procedure is discussed in the paper. Basic results of the research are presented. A classification table of predicted values and factual observed values is shown, the overall percentage of correct recognition is determined. Regression equation coefficients are calculated, the general regression equation is written based on them. Based on the results of logistic regression, ROC analysis was performed, sensitivity and specificity of the model are calculated and ROC curves are constructed. These mathematical techniques allow carrying out diagnostics of health of children providing a high quality of recognition. The results make a significant contribution to the development of evidence-based medicine and have a high practical importance in the professional activity of the author.