Setiadi, Dwi Harid (2018) Pemetaan Kabupaten/Kota Di Jawa Timur Berdasarkan Indikator Daerah Tertinggal Dengan Metode Data Campuran Ensemble ROCK dan SWFM. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
|
Text
06211440000066-Undergraduate_Theses.pdf - Accepted Version Download (1MB) | Preview |
Abstract
Provinsi Jawa Timur merupakan provinsi dengan jumlah Kabupaten/Kota yang termasuk daerah tertinggal terbanyak diantara provinsi di Pulau Jawa.. Penelitian yang dilakukan Kementerian DPDTT terkait penetapan daerah tertinggal hanya mengakomodasi data yang bersifat numerik. Oleh karena itu, dilakukan penelitian dengan metode yang berbeda yaitu dengan memetakan Kabupaten/Kota di Provinsi Jawa Timur yang mengakomodasi data numerik dan kategorik dengan mempertimbangkan aspek statistik. Pada penelitian ini dilakukan analisis pengelompokan menggunakan metode data campuran yaitu ensemble Robust Clustering using Links (ROCK) dan Similarity Weight and Filter Methods (SWFM). Hasil analisis menujukkan metode ensemble ROCK lebih baik karena memiliki akurasi sebesar 0,724 yang lebih besar dari akurasi SWFM yaitu 0,618. Metode ensemble ROCK menghasilkan 4 klaster optimum. Dimana klaster 1 menunjukkan kondisi Kabupaten/Kota tertinggal karena memiliki 11 indikator yang rendah dibanding klaster lainnya. Hasil pengelompokan menunjukkan bahwa 32% Kabupaten/Kota termasuk klaster daerah tertinggal dari total 38 Kabupaten/Kota di Jawa Timur. Selain itu, 75% Kabupaten di Pulau Madura merupakan daerah tertinggal. =============================================================================== East Java Province is a province that has the most underdeveloped districts among provinces in Java. Research conducted by the Ministry of DPDTT concerning the determination of underdeveloped areas only accommodates numerical data. Therefore, the research conducted with different methods that is by mapping the District in East Java Province that accommodates numerical and categorical data by considering the statistical aspects. In this research, grouping using mixed data method of Robust Clustering ensemble using Links (ROCK) and Similarity Weight and Filter Methods (SWFM). The results showed ensemble ROCK method is better because it has an accuracy of 0.724 greater than the accuracy of SWFM is 0.618. The Ensemble ROCK method produces 4 optimum clusters. Where cluster 1 shows cities left behind because it has 11 indicators that are lower than other clusters. The results of the clustering indicate that 32% of districts including clusters of underdeveloped district from a total 38 districts in East Java. In addition, 75% of districts in Madura Island are underdeveloped districts.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Data Campuran, Indikator Daerah Tertinggal, ROCK, SWFM |
| Subjects: | Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics) |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Dwi Harid Setiadi |
| Date Deposited: | 30 Jun 2021 06:33 |
| Last Modified: | 30 Jun 2021 06:33 |
| URI: | https://repository.its.ac.id/id/eprint/56698 |
Actions (login required)
 |
View Item |
