A survey on genetic algorithm based feature selection for disease diagnosis system

Computer Aided Diagnosis (CAD) has been a rapidly growing, dynamic area of research in medical imaging. In recent years, significant and serious efforts have been made towards the development of the CAD system in diagnostic radiology. Machine learning (ML) plays a vital role in CAD because objects such as organs may not be signified precisely by a simple equation and therefore pattern recognition essentially involves learning from examples. In order to reduce the dimensions of the dataset and to increase the classification accuracy rate the feature selection has to be done. Feature Selection (FS) is an important issue in building classification systems. Evolutionary algorithms are an important emergent computing methodology. Genetic Algorithm (GA) being a heuristic search algorithm is generally used to detect important features for large dimensional datasets. This paper surveys the existing literature about the GA for the feature selection. This study also comprises a snapshot of GA from the author's perspective, including variations in the algorithm, modifications and refinements introduced to prevent the local convergence and hybridization of GA with other heuristic algorithms. In the last part of the paper, some of the topics within this field are listed that has to be considered as promising areas of future research.