Breast Cancer Diagnosis Using Feature Ensemble Learning Based on Stacked Sparse Autoencoders and Softmax Regression

Area of Science:

• Medical Informatics
• Machine Learning
• Oncology

Background:

• Breast cancer is the most common cancer in women.
• Early detection significantly improves treatment outcomes.
• Accurate classification of benign versus malignant tumors is crucial.

Purpose of the Study:

• To develop an effective feature ensemble learning model for breast cancer classification.
• To improve the accuracy and performance of early breast cancer detection systems.
• To compare the proposed model against existing state-of-the-art classifiers.

Main Methods:

• Utilized Sparse Autoencoders and Softmax Regression for feature extraction and classification.
• Employed the Breast Cancer Wisconsin (Diagnostic) dataset from the UCI machine learning repository.
• Assessed performance using metrics including accuracy, specificity, sensitivity, precision, and MCC.

Main Results:

• The proposed feature ensemble learning model achieved a high true classification accuracy of 98.60%.
• Demonstrated superior performance compared to Stacked Sparse Autoencoders and Softmax Regression (SSAE-SM) and other classifiers.
• Results indicate the model's efficiency and benefits for breast cancer classification.

Conclusions:

• The proposed model is an efficient and beneficial tool for classifying breast cancer.
• The approach shows promising results and is comparable to existing machine learning methods.
• Feature ensemble learning offers a robust strategy for improving early cancer detection.