Negara, Kadek Adi Surya (2024) Klasifikasi Rumah Tangga Miskin di Provinsi Jawa Timur Menggunakan Metode Support Vector Machine (SVM). Other thesis, Institut Teknologi Sepuluh Nopember.
Text
5003201088-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy |
Abstract
Indonesia masih menghadapi tantangan serius terkait kemiskinan yang disebabkan karena ketidakakuratan metode pengklasifikasian status kemiskinan rumah tangga. Berdasarkan data kemiskinan pada Bulan Maret tahun 2023, Provinsi Jawa Timur menduduki peringkat pertama dengan jumlah penduduk miskin terbanyak di Indonesia. Permasalahan kemiskinan tersebut kemudian diatasi dengan memaksimalkan pengklasifikasian rumah tangga miskin di Provinsi Jawa Timur. Penelitian ini menggunakan data sekunder yang diperoleh dari Susenas yang dilaksanakan oleh BPS Provinsi Jawa Timur pada Bulan Maret tahun 2023. Metode analisis yang digunakan pada penelitian ini adalah Support Vector Machine (SVM) dengan model kernel linier dan kernel Radial Basis Function (RBF) untuk menyelesaikan permasalahan klasifikasi dengan menemukan hyperplane yang memisahkan dua dataset dari dua kelas yang berbeda. Berdasarkan hasil seleksi variabel dengan uji signifikansi regresi logistik biner diperoleh 11 variabel yang berpengaruh dalam melakukan klasifikasi rumah tangga miskin dan rumah tangga tidak miskin. Variabel tersebut diantaranya adalah jumlah ART, status wilayah tempat tinggal, status KRT dalam pekerjaan utama, status literasi KRT, jenis bahan bangunan dinding rumah, luas lantai rumah, status penggunaan jamban, sumber air minum utama, kepemilikan tanah/lahan, kepemilikan sepeda motor, dan kepemilikan komputer/laptop/tablet. Pada Provinsi Jawa Timur, terjadi ketidakseimbangan data, sehingga di atasi dengan metode Synthetic Minority Oversampling Technique (SMOTE). Berdasarkan hasil klasifikasi dengan oversampling 80%, kedua model kernel yang digunakan mampu mengklasifikasikan rumah tangga miskin dan rumah tangga tidak miskin secara seimbang. Model terbaik SVM untuk memodelkan status kemiskinan rumah tangga di Provinsi Jawa Timur adalah menggunakan SVM kernel linier dengan parameter C = 0,1. Hasil ketepatan klasifikasi yang dapat dicapai yakni akurasi sebesar 78,734%, sensitivitas sebesar 77,070%, spesitifisitas sebesar 78,818%, G-Mean sebesar 77,939%, dan AUC sebesar 77,944%.
========================================================================================================================
Indonesia still faces serious poverty-related challenges caused by inaccurate methods of classifying household poverty status. According to March 2023 poverty data, East Java province ranks first with the largest number of poor people in Indonesia. We then addressed the issue of poverty by optimizing the classification of impoverished households in East Java Province. This research uses secondary data obtained from Susenas conducted by BPS East Java Province in March 2023. This study employs the Support Vector Machine (SVM) analytical method, which utilizes a linear kernel model and a Radial Base Function (RBF) kernel to solve the classification problem by identifying a hyperplane that divides two datasets into two distinct classes. The binary logistic regression significance test, based on the variable selection results, yielded 11 variables that significantly influenced the classification of poor households and non-poor families. These variables include the number of ARTs, the status of the area of residence, the KRT status in the main job, the KRT literacy status, the type of building material on the house's walls, the size of the home's floor, the state of use of the dam, the primary drinking water source, land ownership, motorcycle ownership, and computer/laptop/tablet ownership. The Synthetic Minority Oversampling Technique (SMOTE) addressed the data imbalance in the East Java province. According to the classification results with 80% oversampling, the two kernel models used were able to classify poor and non-poor households in a balanced manner. To model the poverty status of households in East Java Province, the best SVM model is to use a linear SVM kernel with a C=0.1 parameter. The classification accuracy achieved was 78.734%, the sensitivity was 77.070%, the specificity was 78.818%, the G-mean was 77.939%, and the AUC was 77.944%.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Klasifikasi, Rumah Tangga Miskin, Support Vector Machine, Classification, Poor Households, Support Vector Machine |
Subjects: | H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation H Social Sciences > HN Social history and conditions. Social problems. Social reform Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA76.6 Computer programming. |
Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
Depositing User: | Kadek Adi Surya Negara |
Date Deposited: | 08 Aug 2024 07:13 |
Last Modified: | 08 Aug 2024 07:13 |
URI: | http://repository.its.ac.id/id/eprint/114618 |
Actions (login required)
View Item |