Text document pre-processing with the KNN for classification using the SVM

Abstract

Document classification can be defined as the task of automatically categorizing collections of electronic documents into their annotated classes, based on their contents. In recent years this has become important due to the advent of large amount of data in digital form. For several decades now document classification in the form of text classification systems have been widely implemented in numerous applications such as spam filtering, e-mails, knowledge repositories and ontology mapping. The main objective is to propose a text classification based on the feature selection and preprocessing there by reducing the dimensionality of the feature vector and increase the classification accuracy. We study the advantages of and disadvantages of K-nearest neighbor (KNN) classification and Support Vector Machine (SVM)classification in performing their classification tasks. In our investigation, we found that the well-performing KNN classification approach may suffer from less accurate than the SVM classification.