The Use of Artificial Intelligence for Automatic Classification of Academic Journals, Office of Academic Resources Center of Nakhon Phanom University

Abstract

     This development of an automated journal classification system using artificial intelligence (AI) for the Library and Information Center, Nakhon Phanom University, utilizes the Dewey Decimal Classification (DDC) system with 20 categories. The data set used for model development consisted of 7,000 journal records, including 2,150 core journal titles obtained from the Office of Academic Resources. The data were divided into a training set (80%) and a testing set (20%).
     The study applied Natural Language Processing (NLP) techniques for text preprocessing, including Thai word segmentation, stop-word removal, and text vectorization. Three AI models were evaluated for journal classification: Naïve Bayes, Support Vector Machine (SVM), and Thai2BERT. Experimental results demonstrated that Thai2BERT achieved the highest performance, with an accuracy of 92.35%, precision of 91.80%, recall of 92.10%, and an F1-Score of 0.919, outperforming both Naïve Bayes and SVM overall.
    The developed system significantly reduced the time required for journal classification and received a high level of user satisfaction. The findings indicate that AI technology can effectively support information resource management and improve the efficiency of automated journal classification services within the Office of Academic Resources.