Cancer Prediction Through a Bacterial Evolutionary Algorithm Based Adaptive Weighted Fuzzy C-Means Approach

Abstract

Clustering of tumor plays a significant part in classifying malignancies from carcinoma genetic data and hence is introduced to deal among the classification problem. It is used in critical applications like cancer treatment for diagnosis and prognosis, analysis of gene expression and related areas. In earlier research works, various tumor clustering schemes were presented based on the single clustering systems and implemented successfully to a variety of biomolecular data for cancer class detection. But, it suffered from some drawbacks similar to starvation of stability, accuracy, and robustness. The ensemble grouping schemes are introduced to overcome these limitations. In this research work, efficient dimensionality reduction is done by using Enhanced Independent Component Analysis (EICA) and effective Feature Selection (FS) using Geometric Particle Swarm Optimization (GPSO) are used. Finally, the competent Adaptive Weighted Fuzzy C-Means Clustering (AWFCM) with metaheuristic optimization scheme of Bacterial Evolutionary Algorithm (BEA) is also proposed. It improves the performance of tumor clustering of biomolecular data. The performance of various ensemble clustering schemes is evaluated based on the efficient feature selection methods.