hMuLab: A Biomedical Hybrid MUlti-LABel Classifier Based on Multiple Linear Regression.

Area of Science:

• Biomedical Informatics
• Machine Learning
• Computational Biology

Background:

• Biomedical classification tasks frequently involve multiple labels, such as genes with diverse functions or patients with comorbidities.
• Existing single-label classification algorithms are insufficient for these complex, multi-label biomedical problems.
• Developing effective multi-label classification methods is crucial for advancing biomedical research.

Purpose of the Study:

• To propose a novel multi-label learning algorithm, named hMuLab, designed for biomedical classification.
• To integrate feature-based and neighbor-based similarity measures within a unified framework.
• To enable accurate assignment of multiple class labels to individual biomedical samples.

Main Methods:

• Developed the hMuLab algorithm, a novel approach for multi-label classification.
• Integrated feature-based and neighbor-based similarity scores to capture sample relationships.
• Utilized multiple linear regression modeling for generating multiple label assignments.
• Evaluated performance using six standard multi-label classification metrics.

Main Results:

• hMuLab demonstrated accurate and stable performance on various biomedical datasets.
• The algorithm effectively handles the inherent multi-label nature of biomedical classification problems.
• Comparative analysis confirmed hMuLab's advantages over existing methods.

Conclusions:

• hMuLab offers a robust solution for multi-label classification in biomedical research.
• The algorithm's ability to assign multiple labels enhances its utility for complex biological data.
• hMuLab represents a valuable addition to the toolkit for analyzing multi-label biomedical datasets.