IMPLEMENTASI ARSITEKTUR AGENTIC AI BERBASIS KNOWLEDGE GRAPH DAN GRAPHRAG UNTUK REKOMENDASI, RISET OBAT HERBAL, DAN EDUKASI INTERAKTIF

Aldida, Jofanza Denis (2026) IMPLEMENTASI ARSITEKTUR AGENTIC AI BERBASIS KNOWLEDGE GRAPH DAN GRAPHRAG UNTUK REKOMENDASI, RISET OBAT HERBAL, DAN EDUKASI INTERAKTIF. S1 thesis, Universitas PGRI Madiun.

	Text Halaman Depan.pdf Download (332kB)
	Text Abstrak.pdf Download (223kB)
	Text Bab I.pdf Download (233kB)
	Text Bab II.pdf Download (1MB)
	Text Bab III.pdf Restricted to Repository staff only Download (383kB)
	Text Bab IV.pdf Restricted to Repository staff only Download (5MB)
	Text Bab V.pdf Download (185kB)
	Text Daftar Pustaka.pdf Download (270kB)
	Text Lampiran.pdf Restricted to Repository staff only Download (969kB)

Abstract

This study aims to design, implement, and test a web-based virtual assistant system that integrates an Agentic AI architecture with a Knowledge Graph (Neo4j) and Graph Retrieval-Augmented Generation (GraphRAG) to address the fragmentation of phytopharmacological information and the risk of medical hallucinations regarding Indonesian herbal medicine. The system development methodology used is Extreme Programming (XP), with visual modeling using the Unified Modeling Language (UML). Data on Indonesian herbal plants including scientific names, active compounds, pharmacological activities, clinical indications, and side-effect risks are structurally mapped into a graph database. In its operation, natural language queries from users are autonomously processed by the Agentic AI component to retrieve contextual facts from the Knowledge Graph via GraphRAG before a local large language model (Qwen LLM) synthesizes the final response. System evaluation was conducted through black-box testing for functional validation, white-box testing using the basis path testing technique to validate internal logical workflows, and user assessment employing a Likert scale. The functional and structural testing results demonstrate that all core features, such as the interactive chatbot service, the educational herbal medicine recommendation system, and the educational quiz, operate validly. Meanwhile, the likert scale assessment yielded a feasibility score of 89.60%, categorizing this system as "Highly Feasible" and reliable for use as a health education and herbal medicine research medium for the general public.

Item Type:	Thesis/Skripsi/Tugas Akhir (S1)
Kata Kunci:	Agentic AI, Knowledge Graph, GraphRAG, Obat Herbal, Edukasi Interaktif. Agentic AI, Knowledge Graph, GraphRAG, Herbal Medicine, Interactive Education.
Subjects:	T Technology > T Technology (General)
Divisions:	Fakultas Teknik > Teknik Informatika
Depositing User:	ALDIDA DENIS JOFANZA
Date Deposited:	30 Jul 2026 06:05
Last Modified:	30 Jul 2026 06:05
URI:	http://eprint.unipma.ac.id/id/eprint/8184

Actions (login required)

View Item