Healthcare Biclustering-Based Prediction on Gene Expression Dataset.

Area of Science:

• Bioinformatics
• Computational Biology
• Health Informatics

Background:

• Gene expression data analysis presents challenges in clustering and information redundancy.
• Healthcare biclustering aims to identify specific gene activities and reduce data complexity.
• Machine learning and heuristic algorithms are increasingly utilized for healthcare biclustering due to their exploration capabilities.

Purpose of the Study:

• To develop an improved healthcare biclustering model for gene expression data.
• To identify specific gene activity patterns and minimize redundant information.
• To evaluate the efficacy of a proposed fuzzy c-means (FCM) clustering method.

Main Methods:

• Development of a novel healthcare biclustering model utilizing fuzzy c-means (FCM) clustering.
• Implementation of two distinct healthcare biclustering approaches for comparative analysis.
• Evaluation based on average match score for overlapping and non-overlapping modules, noise influence, and runtime.

Main Results:

• The proposed FCM clustering method achieved a higher average match score compared to existing PSO-SA and fuzzy logic methods.
• The FCM approach demonstrated reduced runtime in healthcare biclustering tasks.
• The model effectively identified specific gene activity and reduced data duplication.

Conclusions:

• The FCM-based healthcare biclustering model offers enhanced performance for gene expression data analysis.
• This method provides a more efficient and accurate approach to identifying gene expression patterns.
• The findings suggest FCM as a promising technique for complex healthcare data challenges.