Optimal cancer classification of microarray data using different optimization techniques

Abstract

Cancer being one of the most vital diseases in the medical history needs adequate focuses on its cause, symptom and detection. Various algorithms and software have been designed so far to predict the disease at cellular level. The most crucial data for sorting the cancerous tissue is the classification of such tissues based on the gene expression data. Gene expression data consists of high amount of genetic data as compared to the number of data samples. Thus, sample size and dimensions are a major challenge for researchers. In this work, four different types of cancers are analyzed viz., breast cancer, lung cancer, leukemia and colon cancer. The analysis is done using various nature-inspired algorithms like Grasshopper Optimization (GOA), Interval Value Based Particle Swarm Optimization (IVPSO) and Particle Swarm Optimization (PSO). To study the accuracy of the data, five different classifiers are used – Random Forest, K-Nearest Neighborhood (KNN), Neural Network and Support Vector Machine (SVM). The comprehensive data analysis is done with the combination of these five classifiers over various datasets of each of the selected cancer type. After deep analyzing different combinations GOA outperformed for almost each dataset. The research work addresses the issue of dimensionality reduction and efforts towards improving accuracy.