Penerapan Metode Knowledge Graph Embedding Dan Random Forest Dalam Analisis Kelayakan Air

Widodo, Celomitha (2026) Penerapan Metode Knowledge Graph Embedding Dan Random Forest Dalam Analisis Kelayakan Air. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221025-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Kelayakan air merupakan faktor penting bagi kesehatan masyarakat, sehingga dibutuhkan metode analisis yang akurat untuk mengklasifikasikannya. Penelitian ini bertujuan menerapkan Knowledge Graph Embedding (KGE) dan Random Forest untuk menganalisis kelayakan air berdasarkan sembilan parameter kualitas air, yaitu pH, Hardness, Solids, Chloramines, Sulfate, Conductivity, Organic Carbon, Trihalomethanes, dan Turbidity, menggunakan Water Potability Dataset yang berisi 3.276 sampel dengan dua kelas, yaitu layak dan tidak layak konsumsi. Penelitian diawali dengan preprocessing data, meliputi penanganan missing value, penghapusan duplikat, dan normalisasi, dilanjutkan pembangunan Knowledge Graph yang merepresentasikan hubungan antara sampel, parameter dominan, parameter terendah, risiko kesehatan, dan organ tubuh terdampak. Graf tersebut dipelajari menggunakan metode TransH untuk menghasilkan embedding, yang kemudian dihitung dengan cosine similarity dan digabungkan dengan fitur numerik sebagai masukan Random Forest. Model dievaluasi menggunakan Accuracy, Precision, Recall, F1-Score, dan ROC-AUC, dengan hasil berturut-turut 0,7866; 0,7417; 0,6953; 0,7177; dan 0,8718. Hasil ini menunjukkan bahwa integrasi KGE dengan Random Forest dapat memperkaya representasi data kualitas air dan memberikan gambaran hubungan antarparameter yang lebih informatif dibandingkan pendekatan klasifikasi konvensional. Pendekatan ini bermanfaat untuk mengidentifikasi parameter air yang paling berkaitan dengan risiko kesehatan, sehingga pemantauan kualitas air dapat dilakukan secara lebih terarah, sekaligus membuka peluang pengembangan sistem klasifikasi kelayakan air berbasis Knowledge Graph.
======================================================================================================================================
Water potability is an important factor affecting public health, making an accurate analytical method necessary for its classification. This study applies Knowledge Graph Embedding (KGE) and Random Forest to analyze water potability using nine parameters: pH, Hardness, Solids, Chloramines, Sulfate, Conductivity, Organic Carbon,
Trihalomethanes, and Turbidity, drawn from the Water Potability Dataset of 3,276 samples with two classes, potable and non-potable. The research begins with preprocessing, including handling missing values, removing duplicates, and normalization, followed by constructing a Knowledge Graph representing relationships among samples, dominant parameters, lowest parameters, health risks, and affected organs. The graph is learned using TransH to produce embeddings, which are computed using cosine similarity and combined with numerical features as input for Random Forest. The model is evaluated using Accuracy, Precision, Recall, F1-Score, and ROC-AUC, yielding 0.7866, 0.7417, 0.6953, 0.7177, and 0.8718, respectively. These results show that integrating KGE with Random Forest enriches the representation of water quality data and provides a more informative depiction of parameter relationships than conventional classification approaches. This approach helps identify water parameters most closely associated with specific health risks, allowing water quality monitoring to be more targeted, while opening opportunities for further development of Knowledge Graph-based water potability classification systems.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kelayakan air, Knowledge Graph, Knowledge Graph Embedding, TransH, Random Forest, Cosine Similarity, Water potability, Knowledge Graph, Knowledge Graph Embedding, TransH, Random Forest, Cosine Similarity.
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD108 Classification (Theory. Method. Relation to other subjects )
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Celomitha Widodo
Date Deposited:	28 Jul 2026 03:03
Last Modified:	28 Jul 2026 03:03
URI:	http://repository.its.ac.id/id/eprint/138282

Actions (login required)

View Item