Analisis Karakteristik Ruas Jalan Menggunakan Dual Graph Embedding dan Pengelompokan Pada Jaringan Jalan

Fauziyah, Intan (2026) Analisis Karakteristik Ruas Jalan Menggunakan Dual Graph Embedding dan Pengelompokan Pada Jaringan Jalan. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221019-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Jaringan jalan merupakan salah satu infrastruktur penting yang mendukung mobilitas masyarakat dan aktivitas ekonomi, terdiri atas persimpangan yang dihubungkan oleh ruas jalan. Ruas jalan dapat dikelompokkan berdasarkan fungsinya dalam jaringan, namun pengelompokan tersebut umumnya masih didasarkan pada tingkat pelayanan sehingga belum mempertimbangkan posisi dan pola keterhubungan antar ruas jalan. Oleh karena itu, penelitian ini bertujuan mengklasifikasikan jenis jalan berdasarkan karakteristik strukturalnya dengan memanfaatkan representasi jaringan jalan. Pada penelitian ini, jaringan jalan ditransformasikan ke dalam bentuk dual graph, di mana setiap ruas jalan direpresentasikan sebagai sebuah node. Struktur graf yang dihasilkan kemudian direpresentasikan ke dalam ruang vektor menggunakan metode graph embedding dengan algoritma node2vec. Selain itu, dihitung pula skor sentralitas untuk menambah informasi terkait tingkat keterhubungan setiap ruas jalan dalam jaringan. Representasi node2vec dan skor sentralitas selanjutnya digunakan sebagai fitur pada proses klasifikasi melalui tiga skenario, yaitu menggunakan skor sentralitas, embedding node2vec, dan gabungan keduanya. Hasil penelitian menunjukkan bahwa skenario gabungan menggunakan model Random Forest menghasilkan kinerja terbaik dengan akurasi sebesar 95.8% dan Macro F1-score sebesar 89.75%. Analisis karakteristik ruas jalan menunjukkan bahwa skor sentralitas mampu menggambarkan tingkat kepentingan ruas jalan dalam jaringan, di mana jenis jalan yang berperan sebagai jalur utama cenderung memiliki nilai sentralitas yang lebih besar dibandingkan jalan lokal. Selain itu, node2vec mampu merepresentasikan struktural jaringan jalan. Hasil tersebut menunjukkan bahwa skor sentralitas dan embedding node2vec saling melengkapi dalam merepresentasikan karakteristik struktural ruas jalan, sehingga menghasilkan klasifikasi jenis jalan yang lebih baik.
======================================================================================================================================
Road networks are essential transportation infrastructures that support community mobility and economic activities, consisting of intersections connected by road segments. Road segments can be categorized according to their functional roles within the network. However, such classification is generally based on the level of service, without considering the positional and connectivity patterns among road segments. Therefore, this study aims to classify road types based on their structural characteristics by utilizing road network representation. In this study, the road network was transformed into a dual graph, where each road segment was represented as a node. The resulting graph structure was then embedded into a low-dimensional vector space using the Node2Vec graph embedding algorithm. In addition, centrality measures were computed to provide information on the connectivity level of each road segment within the network. The Node2Vec embeddings and centrality measures were subsequently used as input features for the classification process under three scenarios: using centrality measures only, Node2Vec embeddings only, and a combination of both. The results show that the combined-feature scenario using the Random Forest model achieved the best performance, with an accuracy of 95.8% and a Macro F1-score of 89.75%. Analysis of road segment characteristics shows that centrality scores can describe the level of importance of road segments within the network, where road types that serve as main routes tend to have higher centrality values than local roads. Additionally, node2vec is capable of representing the structural characteristics of the road network. These results indicate that centrality scores and node2vec embeddings complement each other in representing the structural characteristics of road segments, thereby resulting in better road type classification.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	jaringan jalan, dual graph, node2vec, sentralitas, klasifikasi, road network, dual graph, node2vec, centrality, classification
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA166 Graph theory
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Intan Fauziyah
Date Deposited:	03 Aug 2026 07:51
Last Modified:	03 Aug 2026 07:51
URI:	http://repository.its.ac.id/id/eprint/142686

Actions (login required)

View Item