Enhanced document classification using noun verb (Nv) terms extraction approach

Abstract

The exponential growth in digital documents and the constantly increasing online information have called for the necessity and lead to classify the documents. Document classification is increasingly vital and indispensable for modern applications. Generally, documents comprise multiple terms of extraction. Here, the main concentration of the most important words is on verbs and nouns, which signify the topics and events. However, nouns and verbs technique or simply called Noun Verb (NV) as an extraction method will greatly enhance the performance of document classification. The aim and the implication of this research is to improve document classification performance by using and utilizing NV extraction to detect the class of a document. Three classifiers namely, K-Nearest Neighbor (KNN), Naive Bayes (NB), and Support Vector Machine (SVM) are used to compare the results. Nine benchmark datasets were employed in testing the proposed document classification. The anticipated classification was verified by evaluating its accuracy. The results exhibit that the verbs as extraction affect document classification. This encouraged the research work to combine verbs with nouns as extraction. The NV method extraction outperformed other extraction methods (e.g., Nouns, Bag of Word (BOW), and Verbs).