Mining in heart rate and sleep databases to discover features relevant to diagnostics of sleep

Abstract

A data mining tool is created to aid the development of new medical diagnostics of sleep. The tool includes WEB accessible database of heterogeneous objective and subjective data on sleep quality of various groups of patients. A toolbox of data mining methods including standard and advanced methods is made available to the users. Preliminary testing of the tool shows that knowledge from the mentioned database can be efficiently elicited by multidimensional scaling, cross-spectrum, and optimization based classification methods, which are relatively new in the data mining field