Evolutionary Optimization Algorithm for Classification of Microarray Datasets with Mayfly and Whale Survival

In the field of bioinformatics, a vast amount of biological data has been generated thanks to the digitalization of high-throughput devices at a reduced cost. Managing such large datasets has become a challenging task for identifying disease-causing genes. Microarray technology enables the simultaneous monitoring of gene expression levels, thereby improving disease diagnosis accuracy for conditions like diabetes, hepatitis, and cancer. As these complex datasets become more accessible, innovative data analytics approaches are necessary to extract meaningful knowledge. Machine learning and data mining techniques can be employed to leverage big and heterogeneous data sources, facilitating biomedical research and healthcare delivery. Data mining has emerged as a vital tool in the medical field, providing insights into illnesses and treatments and enhancing the efficiency of healthcare systems. This thesis aims to present a novel hybrid technique for feature selection using amalgamation wrappers. The proposed approach combines the Mayfly and whale survival strategies, leveraging the strengths of both algorithms. The model was evaluated using various datasets and assessment criteria, including precision, accuracy, recall, F1-score, and specificity. The simulation results demonstrated that the proposed integrated optimization model exhibits improved classification performance with 12% higher accuracy in disease diagnosis.