An Effective Hough Transform and Ensemble Classifier Based Feature Selection and Classification Techniques For Prediction of Breast Cancer

  • Rohit Agarwal et al.

Abstract

Among women, most commonly identified cancer type is breast cancer and major reason for increasing mortality rate among women. As the diagnosis of this disease manually takes long hours and the lesser availability of systems, there is a need to develop the automatic diagnosis system for detection of cancer. Data mining techniques contribute a lot in the development of such system. The major goal of this research work is assessingclassification algorithm’s prediction accuracy with respect to efficiency and effectiveness. An ensemble based classification techniques to predict breast cancer is proposed to enhance the accuracy of classifier. This technique comprises of four phases. Initially in first phase the preprocessing is done and in second phase Otsu’s Gaussian thresholding approach is used for the segmentation of image into objects of background and foreground. Hough transform is used to extract features in third phase. Hough transform is a two dimensional one.Specific shape in an image are isolated by this transform. Finally an assortment of these procedures, including Enhanced Kernel based Support Vector Machine (EKSVM), Enhanced Deep Neural Network and Weighted Random Forest Algorithm has been usually used as an ensemble model. Based on the majority voting the classification model predict the disease effectively. The classifier’s sensitivity, specificity and accuracy can be enhanced by the selection features as shown by results. The better results on Wisconsin breast cancer datasets is produced by the proposed model when comparing with existing models. 

Published
2019-12-12
How to Cite
et al., R. A. (2019). An Effective Hough Transform and Ensemble Classifier Based Feature Selection and Classification Techniques For Prediction of Breast Cancer. International Journal of Control and Automation, 12(6), 26 - 37. Retrieved from https://sersc.org/journals/index.php/IJCA/article/view/2016