Data mining v medicínských datech

In this thesis, medical data have been cleaned, arranged, and analyzed to provide results, which could be further consulted with medical specialists and possibly reflected in medical practice. Specifically, two analytical tasks have been designed. The first one is focused on predicting diseases and revealing disease determinants. The second task is aimed at differences in blood pressure levels of hypertensive patients in dependence on their prescribed medications. The analyzed data have been collected under the aegis of the Research Center EuroMISE -- Cardio within the project of ADAMEK. An overview of eight popular data-mining techniques has been formed. The tasks have been solved using two of those techniques (logistic regression and neural networks; in the first task only) in combination with statistical methods (hypothesis testing). For these purposes, analytical tools by Microsoft (Microsoft SQL Server Analysis Services and Microsoft Office Excel) have been explored and employed. Possible applications of results include improvements in targeting health care and enhancements in systems for early diagnosis, which might result in reducing expenditures and providing patients with a better quality of life.