Applying Cost-Sensitive Extreme Learning Machine and Dissimilarity Integration to Gene Expression Data...

Area of Science:

• Machine Learning
• Bioinformatics
• Computational Biology

Background:

• High-scale, redundant, and imbalanced datasets, such as gene expression data, pose significant challenges for traditional classification methods.
• Existing classification algorithms often incur high costs and lack stability when dealing with such complex data.

Purpose of the Study:

• To develop a cost-sensitive classification algorithm that enhances stability and reduces costs for imbalanced datasets.
• To extend the previously developed Dissimilar ELM (D-ELM) by incorporating misclassification and rejection costs.

Main Methods:

• Introduction of misclassification costs into the Dissimilar ELM framework, creating the cost-sensitive D-ELM (CS-D-ELM).
• Embedding rejection costs into the CS-D-ELM to further improve classification stability.
• Evaluation of the proposed algorithm on relevant datasets, likely including gene expression data.

Main Results:

• The rejection cost-embedded CS-D-ELM algorithm significantly reduces both average and overall classification costs.
• The proposed method maintains competitive classification accuracy despite the cost-reduction focus.
• Demonstrated effectiveness in handling redundant and imbalanced data, exemplified by gene expression data classification.

Conclusions:

• The cost-sensitive D-ELM with embedded rejection costs offers a robust solution for classifying complex, imbalanced datasets.
• The algorithm provides a practical approach to minimize classification expenses while ensuring reliable performance.
• The methodology is adaptable for various classification tasks involving similar data characteristics.