Decision Support for Breast Cancer Detection: Classification Improvement Through Feature Selection.

Area of Science:

• Biomedical Informatics
• Computational Biology
• Medical Data Analysis

Background:

• Early breast cancer detection is crucial for patient outcomes.
• Statistical methods are increasingly used to support medical diagnosis.
• Existing methods for breast cancer classification face challenges with diverse datasets.

Purpose of the Study:

• To develop and evaluate a novel feature selection method for breast cancer classification.
• To improve the accuracy and efficiency of early breast cancer detection using patient data.
• To validate the proposed method on both traditional and novel biological datasets.

Main Methods:

• A feature importance index combining Principal Component Analysis (PCA) and Bhattacharyya distance was developed.
• Iterative classification using k-Nearest Neighbor, linear discriminant analysis, and probabilistic neural network was performed.
• Feature selection was optimized by removing the least important feature after each classification round.

Main Results:

• The method achieved an average accuracy of 99.17% on the Wisconsin Breast Cancer Database (WBCD), retaining 4.61 out of 9 features.
• On protein microarray data, an average accuracy of 98.30% was achieved, retaining 2.17% of the original features.
• The results are comparable to state-of-the-art methods while utilizing simpler, widely available multivariate techniques.

Conclusions:

• The proposed feature selection strategy is effective for early breast cancer detection.
• This method offers a robust and efficient approach for classifying breast cancer cases.
• The findings can significantly aid healthcare professionals in making timely and accurate diagnoses.