The use of data mining methods to identify risks of fetal pathology

Abstract

Listen

Translate article

Reducing child mortality is reflected in several United Nations Sustainable Development Goals and is a key indicator of human progress. The use of information technology makes it possible to process a large number of generated statistical data, or to form new ones over time based on the results of analyses and examinations of a huge number of patients, use the results to determine deviations in the early stages and identify the influence of certain indicators on the likelihood of pathologies. Based on the analysis, the functional requirements for the system are formulated and the task of developing a new information and analytical system is set. The article is devoted to the actual problem of determining the risks of fetal pathology during pregnancy. A variant of the classification of the presence of fetal abnormalities using multinomial logistic regression and a method for clustering patient risk groups using the k -means method was proposed, an optimization technique based on the Broyden - Fletcher - Goldfarb - Shanno method was applied, where memory (L-BFGS) was used as a limitation, the classification of fetal health occurs according to the specified three classes. An information system has been developed to visualize the results of analytical methods. When creating the software package, the client - server architecture was used, which allows the server device to support multi-user mode and provide simultaneous work with several clients. The system includes: an application for Windows, an application for Android. Server interaction module, database interaction module, data analysis module, patient data processing module, doctor data processing module, module, statistical sampling module, personal data protection module, rights differentiation module, visualization module. The modular architecture allows you to upgrade the system, add new modules as needed and make changes to the modules while maintaining the integrity of the system. The developed system has passed the stages of testing and performance evaluation.