Predicting the Cancer Diseases on Microarray Dataset using Simultaneous Feature and Sample Selection with ELM Random Forest Algorithm

Abstract

newline The Microarray technology plays an important role for the biologists to determine the expression rates of thousands of genes. The microarray data consists of small sample and huge dimensional data. In recent days, current research problem faced by the gene expression data classification is that the number of genes greatly exceeds the sample size generally referred with high dimensionality data. In order to avoid the problem of dimensionality, dimension reduction plays a crucial role in DNA microarray analysis. The data mining technique like feature selection plays an important role to solve the problem of dimensionality. The feature selection method differentiates the relevant and irrelevant features and eliminates the irrelevant features. A feature selection strategy using Particle Swarm Optimization (PSO) was introduced to reduce the dimensionality of the microarray data. The reduced features were fed into the different classifiers such as Naive Bayes (NB) and Support Vector Machine (SVM) for classifying the microarray data. At times, the PSO has slow convergence problem. So, to select the feature subsets, a Modified Whale Optimization Algorithm (MWOA) was proposed to select the most relevant features in microarray cancer dataset. The MWOA is a modified version of WOA algorithm which can be employed with fitness function to find the location of the agents at minimum distance. However, the MWOA will get struck into local optima and degrades cancer detection accuracy so some enhanced feature selections have been illustrated briefly in this research work. newlineThe first phase of search work introduces a methodology called search space MWOA (SMWOA) is proposed to enhance the MWOA-based dimensionality of microarray data for cancer detection. To solve the local optima and degrades cancer detection accuracy from the MWOA nonlinear dynamic strategy based on a cosine function was introduced.