Multiple classifier systems for breast mass classification

The American Cancer Society (ACS) recommends women aged 40 and above have a mammogram every year as a Gold Standard for breast cancer detection. Multiple Classifier Technique, which is a hybrid intelligent system, aims to improve the Classification accuracy rate over single classifiers. In this paper, we present an effective approach to breast mammogram analysis to modify the classification accuracy of ensemble neural networkin which we utilize BI-RADS features that were combined with patient´s age and subtlety value, which has been tested on a widely available Digital Database of Screening Mammography (DDSM). In our proposed method, we created an ensemble cluster by using Bagging, AdaBoost, Rotation Forest and reached 92% overall classification accuracy.