Deep Recurrent Speeded Robust Feature Learning Based Bagging Ensemble Multinomial Regressive Cancer Classification Using Mammograms

Breast cancer is one of harmful diseases which impact the living rate of women life. There is currently no proficient way to protect the women life from breast cancer. However prediction of cancer disease in previous stages can raise the possibilities of patient's being recovered and surviving. In existing, various Recurrent Neural Network (RNN) models were designed to perform breast cancer analysis. But, feature extraction performance of conventional RNN is reduced by gradient issues which increase training time, poor cancer classification performance, and also does not give better accuracy for early identification. Therefore, a novel Deep Recurrent Speeded Robust Feature Learning Based Bagging Ensemble Multinomial Regressive Classification (DRSRFL-BEMRC) Model is intended in this paper. The DRSRFL-BEMRC Model initially implements Deep Recurrent Speeded Robust Multiple Feature Extraction (DRSRMFE) algorithm with the goal of extracting relevant features in mammograms with lesser time consumption through operating well to rotation-invariant, robust to noise and computation time and accuracy. Subsequently, Bagging Ensemble Multinomial Regressive Cancer Classifier (BEMRCC) algorithm is constructed to exactly classify an input mammograms as normal, benign or malignant based on the discovered features and thereby lessens false rate of cancer analysis for early recognition. The efficiency of proposed CWDCMFE-MBRC technique is analyzed by taking the factors such as disease classification accuracy, disease classification time, and sensitivity, specificity and false rate depends on dissimilar numbers of input mammograms.