Penerapan Metode Data Campuran Ensemble K-Modes dan Similarity Weight and Filter Method (SWFM) pada Pengelompokan Kabupaten/Kota di Jawa Timur Berdasarkan Indikator Daerah Tertinggal

Paramardini, Larasati Dewinta (2018) Penerapan Metode Data Campuran Ensemble K-Modes dan Similarity Weight and Filter Method (SWFM) pada Pengelompokan Kabupaten/Kota di Jawa Timur Berdasarkan Indikator Daerah Tertinggal. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06211440000058-Undergraduate_Theses.pdf]
Preview
Text
06211440000058-Undergraduate_Theses.pdf - Accepted Version

Download (2MB) | Preview

Abstract

Sebagian besar metode multivariat analisis klaster hanya berfokus pada data skala numerik, sedangkan sering kali data yang dimiliki berupa skala campuran (numerik dan kategorik). Ensemble K-Modes dan Similarity Weight and Filter Method (SWFM) merupakan metode analisis klaster yang dapat menangani data berskala campuran. Keduanya tersusun dari hasil analisis klaster data numerik dan kategorik yang digabungkan. Indikator ketertinggalan daerah di kabupaten/kota Jawa Timur terdiri atas data campuran berskala numerik dan kategorik. Oleh karena itu untuk mengetahui daerah apa saja yang tergolong tertinggal dilakukan pengelompokan dengan metode analisis klaster data campuran. Dengan membandingkan hasil pengelompokan, diperoleh hasil bahwa menggunakan metode ensemble k-modes memiliki nilai akurasi yang lebih besar dibandingkan SWFM yaitu 72.638%. Berdasarkan karakteristik pada tiap klaster terbentuk, klaster 2 yang terdiri dari 39.47% kabupaten/kota di Jawa Timur menunjukkan kondisi yang paling rendah sehingga dianggap sebagai kelompok daerah sangat tertinggal. Daerah lainnya dengan status tertinggal terdapat pada klaster 5 yang terdiri dari 28.95% kabupaten/kota di Jawa Timur.
==================================================================================================================
Most of the multivariate cluster analysis methods only focus on numerical scale data, whereas the data are often in the form of mixed scales (numerical and categorical). Ensemble K-Modes and Similarity Weight and Filter Method (SWFM) is a cluster analysis that could handle mixed data. Both are composed by combined the result of numerical and categorical group data analysis. Underdeveloped district indicator in East Java consists of numerical and categorical mixed data. Therefore, to know the districts of regions are classified as underdeveloped district is done by mixed data clustering analysis method. By comparing the results of clustering, it was found that using k-modes ensemble has a greater accuracy value than SWFM that is 72.638%. Based on the characteristics in each cluster, cluster 2 composed by 39.47% of the district in East Java shows the lowest condition so that it is considered as a very underdeveloped district group. Others area with underdeveloped status are found in cluster 5 that composed by 28.95% of the district in East Java.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Daerah Tertinggal, Ensembel K-Modes, Skala Campuran, 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: Larasati Dewinta Paramardini
Date Deposited: 30 Jun 2021 06:28
Last Modified: 30 Jun 2021 06:28
URI: http://repository.its.ac.id/id/eprint/56585

Actions (login required)

View Item View Item