Breast cancer diagnosis using feature extraction and boosted C5.0 decision tree algorithm with penalty factor

Area of Science:

• Medical Informatics
• Machine Learning in Healthcare
• Bioinformatics

Background:

• Class imbalance is a significant challenge in breast cancer diagnosis, potentially leading to biased classification models.
• Accurate breast cancer diagnosis is crucial for effective treatment and patient outcomes.
• Existing machine learning methods may struggle with imbalanced datasets, impacting diagnostic reliability.

Purpose of the Study:

• To develop a hybrid machine learning approach to address the class imbalance problem in breast cancer diagnosis.
• To enhance the diagnostic accuracy of breast cancer detection using feature extraction and ensemble classification.
• To propose a novel method as an alternative for class imbalance learning in medical applications.

Main Methods:

• A hybrid method combining Principal Component Analysis (PCA) for feature reduction and a boosted C5.0 decision tree algorithm for classification.
• Utilization of a penalty factor to optimize classification performance.
• Implementation and evaluation on biased-representative breast cancer datasets from the UCI machine learning repository.

Main Results:

• The proposed hybrid method effectively handles class imbalance in breast cancer datasets.
• Feature extraction using PCA significantly improved diagnostic accuracy.
• The boosted C5.0 decision tree classifier demonstrated robust performance in classifying breast cancer cases.

Conclusions:

• The hybrid PCA and boosted C5.0 decision tree method is a promising approach for breast cancer diagnosis.
• Feature extraction is essential for achieving high diagnostic accuracy in breast cancer detection.
• This method offers a viable alternative for class imbalance learning in medical contexts.