Classifier ensemble selection based on affinity propagation clustering.

Area of Science:

• Bioinformatics
• Machine Learning
• Data Science

Background:

• High-dimensional data presents challenges for accurate classification.
• Effective feature selection is crucial for improving model performance and interpretability.

Purpose of the Study:

• To propose a novel ensemble feature selection method for high-dimensional data.
• To enhance classification performance using cluster grouping and ranking aggregation techniques.

Main Methods:

• Features correlated with classification were identified using ranking aggregation.
• Affinity propagation clustering with bicor correlation grouped features.
• Ensemble classifiers were trained on diverse feature subsets and selected via kappa coefficient and majority voting.

Main Results:

• The proposed method demonstrated low classification error rates across five gene expression datasets.
• Consistent performance was observed in terms of sensitivity, specificity, accuracy, and G-Mean.
• The method exhibited strong scalability for analyzing large datasets.

Conclusions:

• The ensemble feature selection method effectively identifies relevant features in high-dimensional data.
• The approach offers a robust and scalable solution for classification tasks, particularly in bioinformatics.
• The proposed technique improves classification accuracy and stability.