Lidiawaty, Berlian Rahmy (2025) Pengembangan Traffic Urgency Model untuk Penentuan Prioritas Aduan Teks Lalu Lintas. Doctoral thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 7026221003_Doctoral.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
7026221003_Doctoral.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy |
Abstract
Dalam domain lalu lintas, aduan masyarakat dapat masuk secara masif dan simultan melalui berbagai platform digital, sehingga menimbulkan tantangan dalam menentukan aduan mana yang perlu diprioritaskan. Untuk menjawab tantangan ini, penelitian ini mengembangkan Traffic Urgency Model (TUM), yaitu model prioritas yang menghitung nilai urgensi teks aduan berbasis informasi semantik. Sebanyak 42.251 data dikumpulkan dari platform X melalui interaksi dengan akun @e100ss Surabaya, dilakukan pembersihan data menjadi 23.361, dan 11.933 di
antaranya dikategorikan sebagai aduan yang valid. Selanjutnya, dilakukan ekstraksi entitas menggunakan Named Entity Recognition (NER), yang dikembangkan menjadi NER-TUM dengan akurasi 87,98% untuk menangkap entitas kontekstual dalam teks aduan lalu lintas berbahasa Indonesia. Model ini mengidentifikasi enam konstruk utama penentu urgensi, yaitu Time (waktu), Lcation (lokasi), People(pihak yang terlibat), Condition (kondisi lalu lintas), Object (obyek), dan Comlaint Number (jumlah teks aduan). Entitas yang dihasilkan digunakan untuk mengelompokkan aduan setopik menggunakan Sentence-BERT dengan ambang kemiripan 0,4, serta dikonversi ke dalam variabel numerik yang dinormalisasi dan diformulasikan sebagai konstanta konstruk. Nilai urgensi dihitung dengan mengalikan konstanta konstruk dengan bobot konstruk yang diperoleh melalui metode Analytical Hierarchy Process (AHP), dengan urutan prioritas Condition (0,382), Location (0,202), People (0,146), Time (0,130), serta Object dan Complaint Number (masing-masing 0,07), dan rasio konsistensi sebesar 0,0087. Hasil evaluasi menunjukkan bahwa TUM mampu menentukan urutan prioritas penanganan aduan sesuai dengan preferensi pakar dari Dinas Perhubungan Kota Surabaya dan telah dievaluasi secara konseptual da praktikal. Model ini dapat diimplementasikan dalam sistem purwarupa dan dapat diintegrasikan dalam sistem manajemen aduan melalui tautan: https://github.com/berlianerel/traffic_urgency_model.
=======================================================================================================================================
In the traffic domain, public complaints may enter massively and simultaneously through various digital platforms, creating challenges in determining which complaints should be prioritized. To address this, this study develops the Traffic Urgency Model (TUM), a prioritization model that calculates the urgency value of complaint texts based on semantic information. A total of 42,251 data points were collected from platform X through interactions with the @e100ss Surabaya account, resulting in 23,361 cleaned entries, of which 11,933 were categorized as valid complaints. Entity extraction was then performed using Named Entity Recognition (NER), which was further developed into NER-TUM with an accuracy of 87.98% to capture contextual entities in Indonesian traffic related complaint texts. This model identifies six key constructs for urgency determination: Time, Location, People, Condition, Object, and Complaint Number. The extracted entities were used to group topic-related complaints using Sentence-BERT with a similarity threshold of 0.4, and were converted into normalized numerical variables formulated as construct constants. The urgency score was calculated by multiplying each construct constant by its corresponding weight, which was obtained using the Analytical Hierarchy Process (AHP), with the following priority order: Condition (0.382), Location (0.202), People (0.146), Time (0.130), and both Object and Complaint Number (0.07), with a consistency ratio of 0.0087. Evaluation results show that TUM can determine the priority order of complaint handling in accordance with expert preferences from the Surabaya Department of Transportation, and has been conceptually and practically validated. The model has been implemented in a prototype system and is available for integration into complaint management systems via:
https://github.com/berlianerel/traffic_urgency_model.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | prioritas teks aduan, traffic urgency model, penanganan teks aduan lalu lintas, prioritas teks aduan lalu lintas, complaint text priority, traffic urgency model, handling traffic complaints text, prioritizing traffic complaints text |
| Subjects: | H Social Sciences > HE Transportation and Communications > HE147.6 Transportation--Planning H Social Sciences > HF Commerce > HF5415.52 Consumer complaints. Complaint letters |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 55003-(S3) PhD Thesis |
| Depositing User: | Berlian Rahmy Lidiawaty |
| Date Deposited: | 07 Aug 2025 03:48 |
| Last Modified: | 07 Aug 2025 03:48 |
| URI: | http://repository.its.ac.id/id/eprint/127918 |
Actions (login required)
 |
View Item |