Indonesian Dewey Decimal Classification System Using Support Vector Machine and Neural Network Algorithms

The Dewey Decimal Classification (DDC) system is used to classify books, allowing books to be categorized according to their respective fields in the library. However, knowledge and understanding of the DDC classification system are not widely known among the general public but are primarily known by librarians. Therefore, an automated DDC classification system is needed to simplify the classification system and recognize the different types of books based on their classifications. Given this problem statement, the researchers were interested in studying the DDC classification system using the support vector machine (SVM) and neural network (NN) approaches. This research compares SVM and NN, for algorithms. The study utilizes a dataset of books from one of the universities in Indonesia. Subsequently, the model is tested under various scenarios by varying the dataset, vectorization methods, and classification algorithms. The results indicate that the SVM model with a combination of TF-IDF feature extraction techniques provides the best performance, reaching 0.79. The research also demonstrates that the addition of keyword features to the title can improve performance.