Comparison Analysis of Linear Discriminant Analysis and Cuckoo-Search Algorithm in the Classification of Breast Cancer from Digital Mammograms

Abstract

Objective: Breast cancer is the most common invasive severity which leads to the second primary cause of death<br />among women. The objective of this paper is to propose a computer-aided approach for the breast cancer classification<br />from the digital mammograms. Methods: Designing an effective classification approach will assist in resolving the<br />difficulties in analyzing digital mammograms. The proposed work utilized the Mammogram Image Analysis Society<br />(MIAS) database for the analysis of breast cancer. Five distinct wavelet families are used for extraction of features<br />from the mammograms of MIAS database. These extracted features are statistical in nature and served as input to the<br />Linear Discriminant Analysis (LDA) and Cuckoo-Search Algorithm (CSA) classifiers. Results: Error rate, Sensitivity,<br />Specificity and Accuracy are the performance measures used and the obtained results clearly state that the CSA used<br />as a classifier affords an accuracy of 97.5% while compared with the LDA classifier. Conclusion: The results of<br />comparative performance analysis show that the CSA classifier outperforms the performance of LDA in terms of breast<br />cancer classification.