Large Scale Hierarchical Classification

Abstract

This study elucidates various algorithms used for document or text classification challenge. A sample data is used in this study on which various algorithms like Support Vector Machines (SVM), Naïve Bayes, Neural Networks and K-Nearest Neighbor are used in order to analyze their performances and accuracies. This study tries to identify the limitations and strength of these algorithms on the given sample data that how optimally they can perform classification. Different validations are used in this study to examine the accuracies regarding the classification can be identified. Validations include Split-Validation, X-Validation and Bootstrapping. Different ways and methods are discussed through which classification is made possible in large hierarchy. Finally this study concludes on the basis of results obtained that which machine learning technique or classifier performed excellent on the provided sample data set and achieved higher accuracy as compared to others.