MDPI - Publisher of Open Access Journals

23 pages, 439 KB

Open AccessArticle

Evaluating Proprietary and Open-Weight Large Language Models as Universal Decimal Classification Recommender Systems

by Mladen Borovič, Eftimije Tomovski, Tom Li Dobnik and Sandi Majninger

Appl. Sci. 2025, 15(14), 7666; https://doi.org/10.3390/app15147666 - 8 Jul 2025

Cited by 3 | Viewed by 2457

Manual assignment of Universal Decimal Classification (UDC) codes is time-consuming and inconsistent as digital library collections expand. This study evaluates 17 large language models (LLMs) as UDC classification recommender systems, including ChatGPT variants (GPT-3.5, GPT-4o, and o1-mini), Claude models (3-Haiku and 3.5-Haiku), Gemini series (1.0-Pro, 1.5-Flash, and 2.0-Flash), and Llama, Gemma, Mixtral, and DeepSeek architectures. Models were evaluated zero-shot on 900 English and Slovenian academic theses manually classified by professional librarians. Classification prompts utilized the RISEN framework, with evaluation using Levenshtein and Jaro–Winkler similarity, and a novel adjusted hierarchical similarity metric capturing UDC’s faceted structure. Proprietary systems consistently outperformed open-weight alternatives by 5–10% across metrics. GPT-4o achieved the highest hierarchical alignment, while open-weight models showed progressive improvements but remained behind commercial systems. Performance was comparable between languages, demonstrating robust multilingual capabilities. The results indicate that LLM-powered recommender systems can enhance library classification workflows. Future research incorporating fine-tuning and retrieval-augmented approaches may enable fully automated, high-precision UDC assignment systems. Full article

(This article belongs to the Special Issue Advanced Models and Algorithms for Recommender Systems)

► Show Figures

Figure 1

13 pages, 1712 KB

Open AccessArticle

Analysis of Concentrations of Loans by Using Book Circulation Data in Korea University Library

by Ji-Ann Yang

Publications 2020, 8(4), 53; https://doi.org/10.3390/publications8040053 - 10 Dec 2020

Cited by 4 | Viewed by 5782

Abstract

In this paper, data of almost 8 million loans of books recorded for 15 years by the Korea University Library are analyzed by using big data analytic techniques. During this period, book circulation decreased with an average annual rate of decline of 4.4%. The use factor of books in each Dewey decimal classification (DDC) class was evaluated to measure how efficiently books were used by library users. Loan frequencies of books were analyzed and meaningful results regarding loan concentrations and the half-lives of books were obtained. It was observed that 50% of the total loans in each year were for 20% of all borrowed books in that year. This phenomenon will be called the 20/50 loan rule, and the set of the top 20% most borrowed books, whose cumulative loan frequencies reach 50% of total loans, will be called a core collection. The 20/50 loan rule shows the loan concentration of library books. The extent of loan concentration gets stronger if loans for two or more consecutive years are concerned. It was found that with high probability, books in a core collection at a specific year are also categorized as a core collection in next years. Moreover, books categorized as a core collection in consecutive years have longer half-lives compared with all other circulating books. Full article

► Show Figures

Figure 1

Search Results (2)

Further Information

Guidelines

MDPI Initiatives

Follow MDPI

Saved Queries

Search Filter Reset All

Years

Feature Papers

Subjects

Journals

Article Types

Countries / Regions

Search Results (2)

Further Information

Guidelines

MDPI Initiatives

Follow MDPI