Triple-Technique Diagnosis Using Machine Learned Classifiers

Siddhant Gada, Het Sheth, Meet Chheda, Ashwini Swain, Kriti Srivastava

Abstract: In present scenario, breast cancer has become most common disease among women. Despite the fact, not all public hospitals have the facilities to diagnose breast cancer in India through mammograms. Delaying the diagnosing may increase the chances of cancer to spread throughout the body. Machine learning techniques have been benevolent in the detection and diagnosis of various diseases due to their accurate prediction performance. Various classifiers may provide differently desired accuracies and it is, therefore, exigent to use the most fitting classifier which provides the best accuracy. This paper documents a study of four machine learned classifiers, namely, Multi-Layer Perceptron (MLP), Support Vector Machines (SVM), Logistic Regression (LR) and K-Nearest neighbour (KNN) on the Wisconsin Diagnosis Breast Cancer (WDBC) dataset. The performance of these algorithms has been analysed using classification accuracy and a confusion matrix. We have also introduced an ensemble of the above mentioned classifiers. The results show that the performance of Linear Regression is far superior to other classifiers.

Keywords: Machine Learning, Support Vector Machine, Artificial Neural Network, k Nearest Neighbour, Logistic Regression

How to Cite?: Siddhant Gada, Het Sheth, Meet Chheda, Ashwini Swain, Kriti Srivastava, "Triple-Technique Diagnosis Using Machine Learned Classifiers", Volume 6 Issue 10, October 2017, International Journal of Science and Research (IJSR), Pages: 234-238, https://www.ijsr.net/getabstract.php?paperid=ART20176765, DOI: https://dx.doi.org/10.21275/ART20176765