Presenting a Fast Classifier Based on Unsupervised Learning for Diagnosis Diseases

Abstract

Abstract. From long ago, decision support systems (DSS) as a vital tool in many industrials is considered by decision-makers. These systems can aid managers in making better decisions by collecting and interpreting data. Medical decision support systems (MDSS) have critical role in medical practice. They can help physicians for improving the quality of medical diagnosis. Classifiers as main core of MDSS systems play an important role in improving their performance. This paper presents an unsupervised learning-based real time classifier which is able to perform recognizing medical patterns with proper precision and speed. In the training phase, the proposed classifier is capable to obtain reference models related to classes using synergic clustering technique and finding the frequency of attributes . In order to evaluate efficiency of the proposed classifier, the UCI datasets including breast cancer (WBCD), liver disease (ILPD) and diabetic disease (PID) are applied. The obtained results indicate the effectiveness of the proposed method.