Decision Tree Classification based Decision Support System for Derma Disease

Abstract

The process to utilize, the relevant information or knowledge extracted from large databases, into decision making process is called Data Mining. It is widely used in each sector but especially it helps a lot in health care sector so that complicated disease can be diagnosed easily and accurately. In order to diagnose the disease, a decision support system is proposed based upon decision tree technique so that necessary decision can be made after analyzing the input related to the patients. The classification technique which is used to build this model is decision tree, various decision tree based techniques are explored in this study and measured using various measures like accuracy, sensitivity, specificity, precision, recall, F-measure and ROC area. The Dermatology disease is all about the study related to skin disease which is extremely difficult because all six different categories of these diseases share the similar clinical features. The function tree technique is performing very well with overwhelming experiment results of 100 % accuracy, 100% sensitivity and 100 % specificity. The feature selection methods are applied to increase the quickness of the model. With the help of feature selection methods, all the redundant and unwanted features will get removed and a set of effective features will only be required for the purpose of diagnosis of disease. Best first search and rank search are the most suitable feature selection method which can be applied to strengthen the efficiency of the proposed model for derma diseases.