Population based study on arsenic induced blood samples employing hybrid metaheuristic optimization based ML approach

Long term exposure to arsenic may lead to the development of multi-organ ailments with several non-dermatological, non-malignant outcomes and hence considered as a paradoxical human carcinogen, requires major attention with comprehensive clinical data analysis. An experimental investigation has been pursued to analyze microarray data set and classify biologically arsenic affected genes endorsed by different machine learning approaches. Microarray data consists of a huge number of biologically insignificant genes. Thus it leads to an expensive computation and restrains the classification accuracy. A thorough and critical analysis has been performed to determine a small set of nonredundant genes to classify the arsenic disease. The analysis outcomes could be further put into action to interpret the therapeutic measures to be taken care of the disease. An experimental investigation has been carried out to deal with this population-based learning problem with the objective of classifying a the small set of gene expressions of the disease for particular demography. The analysis has been accomplished with the implementation of a hybrid metaheuristic optimization based machine learning methods for classification of arsenic disease into various classes.