Title: Comparative study of different machine learning classifiers for mammograms and brain MRI images
Authors: Poonam Sonar; Udhav Bhosle; Chandrajit Choudhury
Addresses: Department of Electronics and Telecommunication Engineering, MCT's Rajiv Gandhi Institute of Technology, Mumbai, University of Mumbai, Mumbai, 400053, India ' Department of Electronics and Telecommunication Engineering, MCT's Rajiv Gandhi Institute of Technology, Mumbai, University of Mumbai, Mumbai, 400053, India ' Department of Electronics and Communication Engineering, National Institute of Technology, Silchar, 7888001, India
Abstract: Today, breast cancer in women has become the leading cause of cancer deaths. Mammography has been the most reliable and accurate technique for early and accurate detection of breast cancer. This paper presents machine learning based mammogram classification techniques. The authors propose an improved hybrid KNN-SVM classifier to improve the performance of the expert system. It is based on mapping feature points to kernel space and finds the K nearest neighbours for a given test data point among the training dataset. This narrow down search for support vectors to the more relevant data points. The proposed algorithm is tested on standard MIAS and DDSM mammograms databases and brain MRI database. The results are compared with different machine learning classifiers such as SVM, KNN, Random Forest, C4.5, Logistic Regression, Fisher Discriminant analysis, Naïve Bayesian classifiers. The results show that the performance of the proposed classifier is better compared to the other classifiers.
Keywords: mammograms; texture feature; SVM; support vector machine; KNN; k-nearest neighbour; hybrid SVM_KNN; machine learning classifier.
International Journal of Image Mining, 2018 Vol.3 No.2, pp.152 - 174
Received: 11 May 2017
Accepted: 27 Feb 2018
Published online: 22 Nov 2018 *