Data Preparation by CFS: An Essential Approach for Decision Making Using C 4.5 for Medical Data Mining

Abstract

Trauma has become the leading cause of death in day to day life. Every year millions of people die and many more are handicapped due to various types of accidents caused by Trauma and many people become handicapped for the rest of their lives. It is necessary to develop a tool for predicting and preventing trauma. Reducing mortality rate and increasing the Health awareness is the aim. We have used the data mining process, to extract the useful data from large datasets. Feature subset selection is of immense importance in the field of data mining. The increased dimensionality of data makes testing and training of general classification method difficult. Mining on the reduced set of attributes reduces computation time and also helps to make the patterns easier to understand. The CFS approach for feature selection is proposed. As a part of feature selection step we used filter approach algorithm as random search technique for subset generation, wrapped with different classifiers/ induction algorithm namely decision tree C 4.5, Naive Bayes, as subset evaluating mechanism on standard datasets. It is mandatory to obtain ethical and legal clearance from regional as well as Institutional Ethics Review Board (IERB), before using data mining tools in health care research. We got Ethical clearance from BGS Hospital for using the datasets. These datasets were gathered from the patient files which were recorded in the medical record section of the BGS Hospital Bangalore. Further the relevant attributes identified by proposed filter are validated using classifiers. Experimental results illustrate, employing feature subset selection using proposed filter approach has enhanced classification accuracy. Applying [DM ] techniques to the data brings about very interesting and valuable results. It is concluded that in this case, comparing the result of evaluating the models on test set, decision tree works better than NaiveBayes. In this paper, we have also used WEKA Tool for creating the models.