Федоров Дмитрий Евгеньевич
Patients Similarity Analysis Methods in Clinical Decision Support Systems
Прикладная математика и информатика
In recent decades, medical informatics (MI) became one of the main drivers of evidence-based medicine, and it is almost impossible to imagine modern clinical practice without the use of MI methods. In addition, the widespread Internet Access and modern web technologies open up new opportunities for the development of health care services and collaborative platforms. In this work the field of clinical decision support systems (CDSS) was considered. The goal of the project was to develop prototype of clinical data gathering and visualization web-service based on modern web-technologies. Its architecture allows to deploy and test different algorithms on presented data. For now, visualization techniques intends to use intelligent data analysis methods. In the project we reviewed popular patient similarity measures, clustering algorithms, recommender system algorithms, etc. Also a bunch of open medical data sources was considered. The resulting prototype can be used in new generation of web-interfaces of CDSS and other clinical information systems.