A Data Mining-Based Approach for Cardiovascular Dysautonomias Diagnosis and Treatment

Autonomic nervous system (ANS) is a control system that acts largely unconsciously and regulates bodily functions. An autonomic malfunction can lead to serious problems related to blood pressure, heart, swallowing, breathing and others. A set of dynamic tests are therefore adopted in ANS units to diagnose and treat patients with cardiovascular dysautonomias. These tests generate big amount of data which are very well suited to be processed using data mining techniques. The purpose of this study is to develop a cardiovascular dysautonomias prediction system to identify the appropriate diagnosis and treatment for patients with cardiovascular dysautonomias using a dataset extracted from the ANS unit of the university hospital Avicenne in Morocco. Classification techniques and association rules were used for the diagnosis and treatment stages respectively. In fact, K-nearest neighbors, C4.5 decision tree algorithm, Random forest, Naïve bayes and Support vector machine were applied to generate the diagnosis classification models and Apriori algorithm was used for generating the association rules. The results obtained for each classifier were analyzed and compared to identify the most efficient one.