Enhancing Book Classification Efficiency: A Comparative Analysis of Technology-Driven Approaches

Abstract

As the number of books available in the public domain increases, there is a growing need for inexpensive and efficient classification of these books into categories. While traditional classification systems like the Dewey Decimal Classification and the Library of Congress Classification are still in use today, technology can simplify the process. This paper explores the use of Latent Dirichlet Allocation (LDA), a form of topic modeling, to classify books into categories and compares its performance with other machine learning models. The LDA model achieved an accuracy of 32.29%, while the Naive Bayes, Support Vector Machine, and multinomial Logistic Regression models outperformed it with accuracies of 85.45%, 86.12%, and 85.12%, respectively. This research emphasizes the need for scalable and cost-efficient solutions in categorizing the expanding pool of books within the public domain. Leveraging technology can effectively address this challenge. The comparative analysis of machine learning models offers valuable insights into their potential for enhancing classification accuracy and streamlining the categorization process, thereby opening promising avenues for future development in this field.

Keywords: Public domain, classification, technology, Latent Dirichlet Allocation (LDA), machine learning models.