Nabila, Nasywa Azzah (2026) Klasterisasi Wilayah Terdampak Banjir di Indonesia Menggunakan Metode Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) dan Bayesian Optimization. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (Cover)
22083010051_Cover.pdf Download (2MB) | Preview |
|
|
Text (Bab 1)
22083010051_Bab 1.pdf Download (252kB) | Preview |
|
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Bab 2)
22083010051_Bab 2.pdf Restricted to Repository staff only until 9 July 2029. Download (625kB) |
|
|
Text (Bab 3)
22083010051_Bab 3.pdf Restricted to Repository staff only until 9 July 2029. Download (579kB) |
|
|
Text (Bab 4)
22083010051_Bab 4.pdf Restricted to Repository staff only until 9 July 2029. Download (2MB) |
|
|
Text (Bab 5)
22083010051_Bab 5.pdf Download (234kB) | Preview |
|
|
Text (Daftar Pustaka)
22083010051_Daftar Pustaka.pdf Download (206kB) | Preview |
|
|
Text (Lampiran)
22083010051_Lampiran.pdf Restricted to Repository staff only until 9 July 2029. Download (243kB) |
Abstract
Floods are the most frequently occurring natural disaster in Indonesia. According to data from BNPB, 1.255 flood events were recorded in 2023, increasing to 1.420 events in 2024. This increase in the number of events was followed by a rise in the number of affected people from 3.860.136 to 6.391.056, as well as an increase in the number of damaged units from 1.083.004 to 2.005.567. Furthermore, each province exhibits distinct flood impact characteristics, both in terms of human casualties and the resulting infrastructure damage. If these distinct impact characteristics are not analyzed in depth, resource allocation for flood disaster mitigation efforts risks being suboptimal and mistargeted. Therefore, this study aims to cluster provinces in Indonesia based on flood impact indicators, including the number of victims, damaged houses, and flooded houses. The focus of this study is the application of the Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) method, optimized using Bayesian Optimization to obtain optimal parameter combinations through a probabilistic modeling approach. HDBSCAN is chosen because it is capable of clustering data with varying densities while simultaneously identifying noise automatically. Using the optimal parameters of min_samples = 2 and min_cluster_size = 2, the model achieves a DBCV score of 0,5150 and a DCSI score of 0,7919, resulting in three main clusters and one noise cluster. Cluster 0 consists of 5 provinces, cluster 1 consists of 4 provinces, cluster 2 consists of 21 provinces, and the noise cluster consists of 8 provinces. Furthermore, this study develops a web-based interactive GUI using Streamlit to visualize the clustering results through interactive mapping.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | H Social Sciences > HA Statistics | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Data Science | ||||||||||||
| Depositing User: | Nasywa Azzah Nabila | ||||||||||||
| Date Deposited: | 09 Jul 2026 03:47 | ||||||||||||
| Last Modified: | 09 Jul 2026 03:47 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/54786 |
Actions (login required)
 |
View Item |