Alternative Relative Discrimination Criterion Feature Ranking Technique for Text Classification

The use of text data with high dimensionality affects classifier performance. Therefore, efficient feature selection (FS) is necessary to reduce dimensionality. In classification challenges, FS algorithms based on a ranking approach are employed improve the To rank terms, most algorithms, such as Relative Discrimination Criterion (RDC) and Improved (IRDC), document frequency (DF) term (TF). TF accepts actual values frequently rarely occurring terms used in existing algorithms. However, these focus number rather than category. this research, an alternative method RDC, called Alternative (ARDC) was proposed, which aims accuracy effectiveness RDC ranking. Specifically, ARDC designed identify commonly positive class. results obtained were compared methods, IRDC, standard benchmarking functions Information Gain (IG), Pearson Correlation Coefficient (PCC), ReliefF. experimental reveal that using suggested Reuters21578, 20newsgroup, TDT2 datasets provides better performance precision, recall, f-measure, when employing well-known classifiers multinomial naïve Bayes (MNB), Support Vector Machine (SVM), Multilayer perceptron (MLP), k-nearest neighbor (KNN), decision tree (DT). Another experiment performed validate proposed technique, showcase novelty approach. utilized 20newsgroup dataset Relevant-Based Feature Ranking (RBFR) technique. Naïve (NB), Random Forest (RF) Logistic Regression (LR) demonstrate ARDC.