MLapRVFL: Protein sequence prediction based on Multi-Laplacian Regularized Random Vector Functional Link.

Area of Science:

• Bioinformatics
• Computational Biology
• Machine Learning in Genomics

Background:

• Protein sequence classification is vital for functional annotation and understanding protein interactions.
• High-throughput sequencing generates vast data, increasing the need for accurate protein classification.
• Current machine learning methods struggle with accuracy, generalization, and broad applicability in protein classification.

Purpose of the Study:

• To develop an advanced protein sequence classifier with improved accuracy and generalization.
• To address the limitations of existing machine learning models in protein classification tasks.

Main Methods:

• Proposed a novel classifier: Multi-Laplacian Regularized Random Vector Functional Link (MLapRVFL).
• Integrated Multi-Laplacian and L2,1-norm regularization into the Random Vector Functional Link (RVFL) framework.
• Evaluated MLapRVFL on two standard biological datasets.

Main Results:

• MLapRVFL demonstrated superior predictive performance compared to existing methods.
• The proposed method showed enhanced robustness and accuracy in protein sequence classification.
• Experimental results confirmed improved generalization capabilities of MLapRVFL.

Conclusions:

• MLapRVFL offers significant advancements in protein sequence prediction accuracy and reliability.
• The novel regularization approach effectively improves classification model performance.
• This method contributes to more effective functional annotation and understanding of protein data.