Predictive Modeling of Clinical Data Using Random Forest Algorithm and Soft Computing

Clinical data which includes data of patients and their symptoms is growing largely these days. Detection of a disease in some cases is expensive in terms of money and amount of effort spent. Predictive modeling aids in the early detection of a disease by using health records (HRs). By applying such techniques on an available clinical dataset, a prediction of the current state of a patients disease can be made. The predictive model, in this paper is a classifier, which uses a combination of the random forest algorithm and the genetic algorithm. Each record from the HRs serves as an input to the classifier. The results of classification show that the random forest algorithm and soft computing techniques give better results.