Book Genre Classification Based on Titles and Synopses Using Machine Learning
Year: 2026
Type: conference-paper
Abstract: The appropriate classification of book genres is crucial for readers as it helps them select reading materials that match their interests. However, the process of book genre classification is often conducted manually and subjectively, which can lead to errors. Several previous studies have applied machine learning approaches for book genre classification, however most of them used datasets from school digital libraries or general platforms such as iPusnas, Kaggle, and Goodreads, and did not specifically compare the performance of several machine learning models on Indonesian-language book data based on titles and synopses from local online sources. This study aims to classify Indonesian language book genres into two main categories, namely fiction and nonfiction, and to evaluate the performance of several machine learning models in performing the classification. The data used in this study consist of book titles and synopses obtained from the Gramedia website. The textual data were processed through preprocessing stages and transformed into numerical representations using the Term Frequency-Inverse Document Frequency (TF-IDF) method. Subsequently, the text representations were classified using several machine learning models. The results show that one of the evaluated models achieved the best performance, with an accuracy of 95.6 percent and a high F1-score, as well as stable performance based on cross-validation testing. These findings indicate that machine learning models are effective for text-based book genre classification. This study is expected to serve as a foundation for the development of more relevant and accurate automated book recommendation systems for readers.
Institution: Muhammadiyah University of Surakarta
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7171934597