Optimasi Sistem Rekomendasi Kuliner Indonesia Dengan Evaluasi Retrieval Dan Generasi Pada Integrasi Fuzzy MADM Dan GraphRAG

Waradana, Muhammad Ridho (2026) Optimasi Sistem Rekomendasi Kuliner Indonesia Dengan Evaluasi Retrieval Dan Generasi Pada Integrasi Fuzzy MADM Dan GraphRAG. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6026241018-Master_Thesis.pdf Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Penelitian ini mengembangkan sistem rekomendasi kuliner halal Indonesia menggunakan pendekatan Retrieval Augmented Generation (RAG) berbasis Knowledge Graph (KG) dan Web search Enrichment. Penelitian ini bertujuan mengoptimasi platform Halal Wave yang saat ini menggunakan Fuzzy Multi-Attribute Decision Making (MADM) sebagai algoritma sistem rekomendasi dengan mengintegrasikan teknologi Large Language Model (LLM), RAG, dan KG melalui lima tahapan metodologi yaitu studi literatur, pengumpulan dan pra-pemrosesan data, pembangunan Knowledge Graph dan vector embedding, integrasi sistem rekomendasi dengan RAG dan analisis hasil RAG. Pendekatan ini memungkinkan sistem memberikan informasi kuliner halal yang akurat serta natural dengan memanfaatkan data BPJPH yang terverifikasi halal kemudian memperkaya informasi dan keterkaitan hubungan antar data. Kontribusi penelitian meliputi pengembangan sistem rekomendasi kuliner halal dengan Fuzzy MADM yang sudah ada, terintegrasi dengan LLM, RAG, dan KG. Hasil evaluasi menunjukkan bahwa sistem yang dikembangkan mampu menghasilkan rekomendasi yang sangat relevan dengan skor Relevancy 0.86, Faithfulness 0.76, dan Correctnes 0.50. Meskipun presisi dan recall pengambilan sedikit lebih rendah dibandingkan Fuzzy Search (presisi sebelum 0.47 berbanding sesudah 0.51, dan recall sebelum 0.49 berbanding sesudah 0.46) dengan trade-off peningkatan latensi, sistem ini berhasil memberikan nilai tambah signifikan melalui interaksi natural dan penjelasan kontekstual.
=======================================================================================================================================
This research develops an Indonesian halal culinary recommendation system using the Retrieval Augmented Generation (RAG) approach based on Knowledge Graph (KG) and Web search Enrichment. This research aims to optimize the Halal Wave platform which currently uses Fuzzy Multi-Attribute Decision Making (MADM) as a recommendation system algorithm by integrating Large Language Model (LLM), RAG, and KG technology through five methodological stages: literature study, data collection and pre-processing, Knowledge Graph and vector embedding construction, recommendation system integration with RAG and RAG result analysis. This approach enables the system to provide accurate and natural halal culinary information by utilizing halal-verified BPJPH data and then enriching the information and interrelationships between data. Research contributions include the development of a halal culinary recommendation system with existing Fuzzy MADM, integrated with LLM, RAG, and KG. The evaluation results show that the developed system is able to produce highly relevant recommendations with a Relevancy score of 0.86, Faithfulness of 0.76, and Correctness of 0.50. Although the retrieval precision and recall are slightly lower than the Fuzzy Search (precision before 0.47 compared to after 0.51, and recall before 0.49 compared to after 0.46) with the trade-off of increased latency, this system succeeds in providing significant added value through natural interactions and contextual explanations.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Retrieval Augmented Generation, Knowledge Graph, Fuzzy, Sistem Rekomendasi, Halal, Recommendation System
Subjects:	Q Science > QA Mathematics > QA166 Graph theory Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.9.I58 Recommender systems (Information filtering) Q Science > QA Mathematics > QA9.64 Fuzzy logic Z Bibliography. Library Science. Information Resources > ZA Information resources > Z699.5 Information storage and retrieval systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis
Depositing User:	Muhammad Ridho Waradana
Date Deposited:	23 Jan 2026 08:47
Last Modified:	23 Jan 2026 08:47
URI:	http://repository.its.ac.id/id/eprint/130247

Actions (login required)

View Item