Sequential pattern recognition for medical records analysis
Andrzej Tolarczyk, Krzysztof Siwek · 2016
Automatic recognition of patterns is an important problem in a variety of engineering and scientific disciplines. The most common application in medicine relates to medical imaging for diagnostics' purpose. This paper presents another field of pattern recognition application - mining medical records. Analysis of treatment pathways (in which diagnostic and therapeutic procedures are sequenced on a timeline) as a mean of providing and standardized patient care of high quality, is one of the issues gaining in importance in the last decade in the field of medical informatics. As there is a need for scalable computational tools for pattern exploration, we propose a conceptual framework, adapted to Poland's context, for expressing data mining tasks to discover knowledge about patients and providers similarities involving multiple time and procedure granularities.