ECG Knowledge Discovery Based on Ontologies and Rules Learning for the Support of Personalized Medical Decision Making

The electrocardiogram (ECG) is the most common non-invasive used method for monitoring the heart condition. Combined with other available patient medical data, the manual analysis and evaluation of ECG data become labour intensive and prone to errors. With the increased amount of available digital medical data, there is a need for proper methods to support medical practitioners in the decision-making process. These practitioners base their diagnostic decisions on standardized procedures, combined with field experience. In this paper, we present the conceptual design for an approach to knowledge discovery of ECG data based on ontologies and rules learning using Learning Classifier Systems (LCS). Ontologies can provide a platform and application-independent representation of knowledge based on the patient's medical data. Furthermore, rule-based reasoning provides a mechanism for discovering new knowledge. LCS provide a tool for automatically discovering new rules that are maximally general and can support sequential decision-making process. The use of LCS and rule-based reasoning provide a mechanism for encoding existing and new knowledge that can improve the efficiency of personalized medical treatment.