Perbandingan Model Sentence Transformer untuk Rekomendasi Game Berbasis Konten dengan Cosine Similarity

Ajeng Miranti, Gina Purnama Insany

Abstract


Penelitian ini membandingkan performa tiga model Sentence Transformer, yaitu all-MiniLM-L6-v2, all-mpnet-base-v2, dan msmarco-distilbert-base-tas-b, pada sistem rekomendasi video game berbasis content-based filtering. Dataset Steam diolah dengan menggabungkan short description, detailed description, dan categories menjadi dokumen teks. Sebanyak 10.000 game dipilih secara acak dan dievaluasi pada 300 query valid. Setiap model menghasilkan embedding, kemudian cosine similarity digunakan untuk membentuk rekomendasi Top-10. TF-IDF digunakan sebagai baseline, sedangkan relevansi rekomendasi ditentukan melalui kesamaan genre dan nilai Jaccard similarity minimal 0,30. Evaluasi menggunakan Precision@10, Recall@10, MAP@10, MRR@10, NDCG@10, serta waktu pembentukan representasi. Hasil menunjukkan bahwa msmarco-distilbert-base-tas-b menghasilkan nilai tertinggi pada Precision@10 sebesar 0,6283, MAP@10 sebesar 0,5036, MRR@10 sebesar 0,7938, dan NDCG@10 sebesar 0,6392. all-MiniLM-L6-v2 menghasilkan kualitas yang mendekati model terbaik dengan waktu representasi 30,73 detik, lebih cepat daripada TAS-B yang memerlukan 129,05 detik. TF-IDF menjadi metode tercepat dengan waktu 1,24 detik, tetapi memiliki kualitas ranking terendah. Hasil ini menunjukkan bahwa pemilihan model perlu mempertimbangkan keseimbangan antara kualitas rekomendasi dan efisiensi komputasi.

Keywords


content-based filtering; cosine similarity; Sentence Transformer; Steam; video game recommendation

References


E. Salkanović, N. Zukorlić, L. Oković, and D. Kečo, “A Data Analysis of Steam’s Game Catalog and Diverse Recommendation Strategies,” International Journal of Computer Applications, vol. 186, no. 54, pp. 39–49, Dec. 2024.

F. Y. Debbek, F. O. Ehtiba, and H. S. B. Abdelmula, “Movie Recommendation Engine Based on Cosine Similarity and KNN,” International Journal of Engineering and Information Technology, vol. 12, no. 1, pp. 266–270, Dec. 2024.

M. Guesmi, M. A. Chatti, L. Kadhim, S. Joarder, and Q. U. Ain, “Semantic Interest Modeling and Content-Based Scientific Publication Recommendation Using Word Embeddings and Sentence Encoders,” Multimodal Technologies and Interaction, vol. 7, no. 91, 2023, doi: 10.3390/mti7090091.

M. T. Colangelo, M. Meleti, S. Guizzardi, E. Calciolari, and C. Galli, “A Comparative Analysis of Sentence Transformer Models for Automated Journal Recommendation Using PubMed Metadata,” Big Data and Cognitive Computing, vol. 9, no. 67, 2025, doi: 10.3390/bdcc9030067.

N. Reimers and I. Gurevych, “Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks,” in Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing, 2019, pp. 3982–3992, doi: 10.18653/v1/D19-1410.

A. P. Putra, D. P. S. Putri, and A. K. A. C. Wiranatha, “Scientific Paper Recommendation System: Application of Sentence Transformers and Cosine Similarity Using arXiv Data,” Journal of Applied Informatics and Computing, vol. 9, no. 4, pp. 1374–1382, Aug. 2025.

P. K. Goyal, S. Singh, and U. S. Tiwary, “silp_nlp at SemEval-2025 Task 5: Subject Recommendation With Sentence Transformer,” in Proceedings of the 19th International Workshop on Semantic Evaluation, 2025, pp. 2455–2460.

M. Rawat, K. Aditya, S. Parasar, and Y. Tuteja, “Enhancing NFT Discovery with Sentence Transformers and Cosine Similarity: A Deep Learning Approach,” International Journal of Modern Science and Research Technology, vol. 3, no. 3, pp. 355–362, Mar. 2025, doi: 10.5281/zenodo.15119700.

L. Yang, Z. Liu, Y. Wang, C. Wang, Z. Fan, and P. S. Yu, “Large-scale Personalized Video Game Recommendation via Social-aware Contextualized Graph Neural Network,” in Proceedings of the ACM Web Conference 2022, 2022, pp. 3376–3386, doi: 10.1145/3485447.3512273.

R. Wu, “RecBERT: Semantic recommendation engine with large language model enhanced query segmentation for k-nearest neighbors ranking retrieval,” Intelligent and Converged Networks, vol. 5, no. 1, pp. 42–52, 2024, doi: 10.23919/ICN.2024.0004.

J. Li, M. Wang, J. Li, J. Fu, X. Shen, J. Shang, and J. McAuley, “Text Is All You Need: Learning Language Representations for Sequential Recommendation,” in Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023, pp. 1258–1267, doi: 10.1145/3580305.3599519.

V. Vančura, P. Kordík, and M. Straka, “beeFormer: Bridging the Gap Between Semantic and Interaction Similarity in Recommender Systems,” in Proceedings of the 18th ACM Conference on Recommender Systems, 2024, pp. 1–6, doi: 10.1145/3640457.3691707.

F. F. Ijebu, Y. Liu, C. Sun, and P. U. Usip, “Soft cosine and extended cosine adaptation for pre-trained language model semantic vector analysis,” Applied Soft Computing, vol. 169, art. 112551, 2025, doi: 10.1016/j.asoc.2024.112551.

FronkonGames, “steam-games-dataset,” Hugging Face Datasets. [Online]. Available: https://huggingface.co/datasets/FronkonGames/steam-games-dataset. [Accessed: Jun. 26, 2026].

P. Jaccard, “Étude comparative de la distribution florale dans une portion des Alpes et du Jura,” Bulletin de la Société Vaudoise des Sciences Naturelles, vol. 37, no. 142, pp. 547–579, 1901, doi: 10.5169/seals-266450.


Refbacks

  • There are currently no refbacks.


Editorial Office :

Prosiding SENDIKO (Seminar Nasional Hasil Penelitian & Pengabdian Masyarakat Bidang Ilmu Komputer)
Published by Universitas PGRI Madiun
Managed by Program Studi Sistem Informasi Fakultas Teknik Universitas PGRI Madiun
Address Jl. Auri 14-16 Kota Madiun Kampus III Universitas PGRI Madiun 63118
Website http://prosiding.unipma.ac.id/index.php/sendiko/index
Email [email protected]

e-ISSN:  3025-4604