Klasterisasi Provinsi di Indonesia Berdasarkan Ketimpangan Pembangunan Ekonomi dengan Pendekatan Algoritma K-Means dan DBSCAN

Tsabita, Hilma (2024) Klasterisasi Provinsi di Indonesia Berdasarkan Ketimpangan Pembangunan Ekonomi dengan Pendekatan Algoritma K-Means dan DBSCAN. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5006201086-Undergraduate_Thesis.pdf] Text
5006201086-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (2MB) | Request a copy

Abstract

Ketimpangan pembangunan ekonomi antarwilayah masih menjadi isu strategis dalam pembangunan nasional Indonesia. Data menunjukkan kesenjangan signifikan dalam hal pendapatan (PDRB) dan persentase penduduk miskin antarprovinsi di Indonesia. Misalnya, PDRB per kapita DKI Jakarta 13-14 kali lipat lebih tinggi daripada Nusa Tenggara Timur
pada tahun 2022. Sementara, persentase penduduk miskin di Papua dan Papua Barat mencapai 20-26% pada tahun 2022, jauh melampaui rata-rata nasional 9,36%. Ketimpangan pembangunan ini berpotensi melahirkan dampak negatif seperti konflik sosial dan melebarnya kesenjangan pembangunan ekonomi antarwilayah. Oleh karena itu, identifikasi pola ketimpangan antarwilayah sangat diperlukan agar intervensi kebijakan afirmatif dapat dirancang lebih spesifik dan tepat sasaran. Penelitian ini bertujuan melakukan klasterisasi 34 provinsi di Indonesia ke dalam beberapa kelompok berdasarkan karakteristik serupa. Klasterisasi dilakukan menggunakan 2 algoritma clustering, yaitu K-means dan Density Based Spatial Clustering of Applications with Noise (DBSCAN). K-means clustering dipilih karena kemampuannya melakukan klasterisasi data dalam jumlah besar secara efisien. Sementara DBSCAN mampu mengelompokkan data yang memiliki kepadatan tertentu dan mengidentifikasi data pencilan. Variabel yang digunakan dalam klasterisasi meliputi PDRB per kapita, persentase penduduk miskin, jumlah penduduk, IPM, dan tingkat pengangguran
terbuka (TPT). Data bersumber dari publikasi resmi BPS tahun 2023. Hasil penelitian menunjukkan bahwa K-Means clustering lebih unggul dalam mengklasterisasi provinsi di Indonesia dengan kombinasi variabel PDRB per kapita, Persentase Penduduk Miskin, Jumlah Penduduk, IPM, dan Tingkat Pengangguran Terbuka (TPT) berdasarkan metrik evaluasi Daevis Bouldin Index (DBI) dan Calinski Harabasz Index (CHI). Hasil ini konsisten dengan perhitungan Indeks Williamson yang mengindikasikan adanya ketimpangan pembangunan ekonomi di Indonesia.
========================================================================================================================
Regional economic development inequality remains a strategic issue in Indonesia's national development. Data shows a significant gap in Gross Domestic Product (GDP) and poverty
rates across provinces in Indonesia. For example, the per capita GDP of DKI Jakarta was 13-14 times higher than that of East Nusa Tenggara in 2022. Meanwhile, the poverty rates in Papua and West Papua reached 20-26% in 2022, far exceeding the national average of 9.36%. This development inequality has the potential to create negative impacts such as social conflicts and widening economic development gaps between regions. Therefore, the identification of inter-regional inequality patterns is crucial so that affirmative policy interventions can be designed more specifically and targeted. This research aims to cluster the 34 provinces in Indonesia into several groups based on similar characteristics. The clustering is carried out using two clustering algorithms, namely K-means and Density Based Spatial Clustering of Applications with Noise (DBSCAN). K-means clustering is chosen for its ability to efficiently Cluster large amounts of data. Meanwhile, DBSCAN is able to group data that has a certain density and identify outlier data. The variables used in the clustering include per capita GDP, poverty rate, population, HDI, and open unemployment rate (TPT). The data is sourced from official BPS publications in 2023. The research results demonstrate the superiority of K-Means clustering in grouping Indonesian provinces based on a combination of variables including GDP per capita, Percentage of Poor Population, Total Population,
Human Development Index (HDI), and Open Unemployment Rate. This superiority is indicated by the evaluation metrics of Davies-Bouldin Index (DBI) and Calinski-Harabasz Index (CHI). These findings are consistent with the Williamson Index calculation, which indicates the existence of economic development disparities in Indonesia.

Item Type: Thesis (Other)
Uncontrolled Keywords: Clustering, DBSCAN, Ketimpangan Pembangunan Ekonomi, K-Means, Wilayah, Clustering, DBSCAN, Economic Development Inequality, K-Means, Regions
Subjects: Q Science > QA Mathematics > QA278.55 Cluster analysis
Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User: Hilma Tsabita
Date Deposited: 31 Jul 2024 08:30
Last Modified: 31 Jul 2024 08:30
URI: http://repository.its.ac.id/id/eprint/110521

Actions (login required)

View Item View Item