Microarray Cancer Classification with Stacked Classifier in Machine Learning Integrated Grid L1-Regulated Feature Selection

Abstract

Cancer is a group of diseases leads to higher mortality rate due to abnormal growth of cell within body tissues. Microarray dataset incorporates features those are challenge due to higher inherent complexity within the data. Through incorporation of the microarray dataset, different characteristics disciplines are incorporated for the different applications. The microarray dataset incorporates the higher dimensionality of data, size of sample, variance in features and imbalance in the class. To overcome the issues associated with the microarray dataset feature reduction and technique for the dimensionality reduction are implemented for the diagnosis and classification. To achieve the effective classification in the microarray dataset Stacked Machine Learning Classifier (SMLC) is developed. The developed SMLC model incorporates the different classifier those are stacked within the Machine learning model. The stacked classifier comprises of the k-nearest classifier, Random Forest and logistics classifier for the classification with the L1-regularization. Through the ranking based approach feature selection is performed within the microarray datasets. The performance evaluation of the proposed SMLC model exhibits the higher classification accuracy of the 99%. Additionally, the proposed SMLC model achieves the 97% of the precision, recall and F-measure.