Title: Machine learning capabilities in medical diagnosis applications: computational results for hepatitis disease
Authors: Joydip Dhar; Ashok Ranganathan
Addresses: ABV-Indian Institute of Information Technology and Management, Gwalior 474015, Madhya Pradesh, India ' ABV-Indian Institute of Information Technology and Management, Gwalior 474015, Madhya Pradesh, India
Abstract: The main goal of the research work is to apply a Genetic Algorithm (GA) in order to prune the inputs for an Artificial Neural Network (ANN) for medical diagnosis in order to reduce the computational complexity. The inputs in medical diagnosis are the diagnostic factors. The GA implemented creates the essential and minimal subset of diagnostic factors required for medical diagnosis. Firstly, the ANN is applied alone and the time taken and efficiency of the medical diagnostic system are recorded. Then, pruning of inputs using GA and then the pruned inputs are used for the ANN, and the time taken and efficiency obtained are compared with the previous one. The medical diagnostic data set is taken from UCI medical repository for the hepatitis disease. There is a significant percentage reduction in training time as well as testing time of ANN and a significant improvement in the success rate of diagnosis.
Keywords: machine learning; genetic algorithms; ANNs; optimisation; artificial neural networks; medical diagnosis; hepatitis; computational complexity.
DOI: 10.1504/IJBET.2015.069398
International Journal of Biomedical Engineering and Technology, 2015 Vol.17 No.4, pp.330 - 340
Received: 19 Oct 2013
Accepted: 30 Apr 2014
Published online: 14 May 2015 *