Efficient Classifier Algorithm for Gene Expression Data Analysis

Abstract

An efficient exploration of likelihood hidden knowledge that are existing over the data is
employed.The preprocessing step is exploited for unwanted data elimination and to reduce
the time taken for computation. The feature selection is made for the selection of more
appropriate features which is accomplished with the use of PCA approach. After the
extraction of best features, classification approach is carried out. In the data analysis of
microarray gene expression, two feature extraction methods are applied like ranking-based
and set-based feature extractions. So as to enhance the efficiency of extraction, a Multi
Algorithm Fusion (MAF) method is instigated. Likewise, the Polynomial Support Vector
Machine (PSVM) that is related to MAF is taken for the process of feature extraction. From
this, the absolute weight of SVM, fisher ratio and PSVM are attained. The MAF dependent
PSVM algorithm is selected for attaining effectual results. It achieved an effective feature
robustness with enhanced MAF-PSVM than the traditional methods and the classification
error was lesser for our proposed approach. Finally, random forest classifier was utilized for
the effective classification approach and it was proved with efficient outcome of measures
such as error rate, accuracy, specificity, precision, recall, F1score, false positive rate.