ACO Based Feature Selection : An Application for Medical Diagnosis

Abstract

Medical Diagnosis is a system for the examination of a man's symptoms based on disorder. This issue has been examined and applied to several healthcare systems in medication. It has substantial fascination in the area of computer science because of huge cause for various ailments. The medical dataset contains immense number of immaterial and repetitive features. Not all the features are required to analyze whether the specific patient is having that specific sickness or not. Feature Selection (FS) is the system for finding the most essential features for a predictive model. This system is used to discover and eliminate not required, insignificant and repetitive features that do not contribute or diminish the accuracy of the predictive model. In this paper, we address the medical diagnosis by utilizing feature selection of Ant Colony Optimization (ACO). It is a nature inspired heuristic algorithm. By utilizing this algorithm, we will attempt the medical diagnosis using FS. Increment in features may lead to decrease in accuracy of the model. In order to attain the increase in accuracy of the model feature selection is used.