Pengembangan Sistem Otomatisasi Manajemen Koleksi dan Pemasaran Digital Perpustakaan ITS Berbasis Zero-Shot Learning dan Large Language Models

Arianto, Muhammad Althaf (2026) Pengembangan Sistem Otomatisasi Manajemen Koleksi dan Pemasaran Digital Perpustakaan ITS Berbasis Zero-Shot Learning dan Large Language Models. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025221225-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Perpustakaan Institut Teknologi Sepuluh Nopember (ITS) menghadapi tantangan inefisiensi operasional akibat proses klasifikasi koleksi digital yang masih manual serta keterbatasan sumber daya dalam menyusun konten pemasaran digital yang efektif. Tugas akhir ini mengembangkan sistem otomatisasi berbasis web yang mengintegrasikan teknologi Zero-Shot Learning (ZSL) dan Large Language Models (LLM) untuk menangani dua permasalahan tersebut secara sekaligus. Pada modul klasifikasi, tiga model Natural Language Inference (NLI) yaitu DeBERTa-v3-Base, DeBERTa-v3-Large, dan BART-Large-MNLI dibandingkan menggunakan 500 artikel akademik dalam 10 kategori Library of Congress Classification (LCC). Setelah dilakukan perbandingan, model DeBERTa-v3-Base terpilih sebagai model terbaik dengan akurasi 82,00% dan F1-Score makro 82,06%. Sementara itu pada modul pemasaran digital, sistem memanfaatkan Groq API untuk menghasilkan konten promosi yang disesuaikan untuk Instagram, WhatsApp, dan TikTok secara otomatis. Pengujian fungsional dilakukan melalui tahapan Black Box Testing terhadap 18 skenario dan seluruh uji coba skenario menunjukkan hasil yang sesuai harapan, dan evaluasi kualitas konten dilakukan melalui Human Evaluation oleh pustakawan Perpustakaan ITS. Hasil evaluasi pengguna oleh empat pustakawan Perpustakaan ITS menunjukkan rata rata penilaian keseluruhan 4,43 dari skala 5, dengan skor tertinggi pada aspek penghematan waktu kerja klasifikasi koleksi, sehingga sistem yang dikembangkan terbukti berpotensi mempercepat pengelolaan koleksi digital sekaligus membantu pustakawan dalam pembuatan konten promosi perpustakaan.
=====================================================================================================================================
The Institut Teknologi Sepuluh Nopember (ITS) Library faces operational inefficiency challenges due to the manual classification process of digital collections and limited resources in developing effective digital marketing content. This final project develops a web-based automation system that integrates Zero-Shot Learning (ZSL) and Large Language Models (LLM) technologies to address these two issues simultaneously. In the classification module, three Natural Language Inference (NLI) models namely DeBERTa-v3-Base, DeBERTa-v3-Large, and BART-Large-MNLI are compared using 500 academic articles in 10 Library of Congress Classification (LCC) categories. After the comparison, the DeBERTa-v3-Base model was selected as the best model with an accuracy of 82.00% and a macro F1-Score of 82.06%. Meanwhile, in the digital marketing module, the system utilizes the Groq API to automatically generate customized promotional content for Instagram, WhatsApp, and TikTok. Functional testing was conducted through black box testing on 18 scenarios, with all scenarios yielding results as expected. Content quality was evaluated through human evaluation by ITS Library librarians. User evaluation results from four librarians at the ITS Library show an overall average rating of 4.43 out of 5, with the highest score recorded for the time savings aspect in collection classification, indicating that the developed system has strong potential to accelerate digital collection management while assisting librarians in creating promotional content for the library.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Large Language Models, Manajemen Koleksi, Otomatisasi Perpustakaan, Pemasaran Digital, Zero-Shot Learning, Collection Management, Digital Marketing, Large Language Models, Library Automation, Zero-Shot Learning.
Subjects:	Q Science > QA Mathematics > QA76 Computer software Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science Z Bibliography. Library Science. Information Resources > Z666.7 Metadata.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Althaf Arianto
Date Deposited:	24 Jul 2026 08:01
Last Modified:	24 Jul 2026 08:01
URI:	http://repository.its.ac.id/id/eprint/137955

Actions (login required)

View Item