Perbandingan Pemodelan Regresi Probit Ordinal Tanpa dan dengan Pendekatan Synthetic Minority Oversampling Technique (Studi Kasus: Status Ketahanan Pangan di Kawasan Timur Indonesia)

Ainin, Muharinda Sugma Nur (2024) Perbandingan Pemodelan Regresi Probit Ordinal Tanpa dan dengan Pendekatan Synthetic Minority Oversampling Technique (Studi Kasus: Status Ketahanan Pangan di Kawasan Timur Indonesia). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5003201017-Undergraduate_Thesis.pdf] Text
5003201017-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (2MB) | Request a copy

Abstract

Pangan merupakan salah satu kebutuhan pokok manusia yang harus dipenuhi setiap saat. Pemerintah Indonesia mengatur penyelenggaraan dan pemenuhan pangan berdasarkan kedaulatan pangan, kemandirian pangan, dan ketahanan pangan melalui Undang-Undang No. 18 Tahun 2012. Ketahanan pangan termasuk salah satu tujuan dalam Sustainable Development Goals (SDGs). Di Indonesia, ketahanan pangan menjadi salah satu isu strategis pembangunan nasional. Berdasarkan Peta Ketahanan dan Kerentanan Pangan (FSVA) Indonesia Tahun 2022 menunjukkan bahwa masih terjadi ketimpangan status ketahanan pangan di beberapa wilayah Indonesia dan sebagian besar berada di Kawasan Timur Indonesia, dimana 25 kabupaten/kota dengan nilai IKP terendah se-Indonesia berada di Kawasan Timur Indonesia. Status ketahanan pangan memiliki skala ordinal dengan kategori 1 merupakan terendah dan kategori 6 merupakan kategori tertinggi. Pada penelitian ini dilakukan pemodelan terkait status ketahanan pangan pada 176 kabupaten/kota di Kawasan Timur Indonesia menggunakan regresi probit ordinal. Berdasarkan data FSVA 2022, ditemukan adanya indikasi data imbalance pada persebaran kategori status ketahanan pangan di Kawasan Timur Indonesia dengan persentase data minoritas sebesar 14,77% sehingga dilakukan pemodelan tanpa dan dengan pendekatan SMOTE. Model terbaik merupakan pemodelan menggunakan regresi probit ordinal dengan pendekatan SMOTE. Hasil dari penelitian ini diperoleh variabel prediktor yang berpengaruh signifikan terhadap model adalah produktivitas tanaman padi, luas lahan panen padi, indeks daya beli masyarakat, persentase rumah tangga penerima bantuan pangan dan persentase bayi BBLR. Pemodelan regresi probit ordinal dengan pendekatan SMOTE memiliki ketepatan klasifikasi sebesar 74,51%.
=========================================================================
Food is one of the basic human needs that must be met at all times. The Indonesian government regulates the implementation and fulfillment of food based on food sovereignty, food independence, and food security through Law No. 18 of 2012. Food security is one of the goals in the Sustainable Development Goals (SDGs). In Indonesia, food security is one of the strategic issues of national development. Based on the 2022 Indonesian Food Security and Vulnerability Map (FSVA), it shows that there is still inequality in food security status in several regions of Indonesia and most of them are in Eastern Indonesia, where 25 districts/cities with the lowest IKP values in Indonesia are in Eastern Indonesia. Food security status has an ordinal scale with category 1 being the lowest and category 6 being the highest. In this study, modeling was carried out related to food security status in 176 districts/cities in Eastern Indonesia using ordinal probit regression. Based on FSVA 2022 data, there was an indication of data imbalance in the distribution of food security status categories in Eastern Indonesia with a percentage of minority data of 14.77% so that modeling was carried out without and with the SMOTE approach. The best model is modeling using ordinal probit regression with the SMOTE approach. The results of this study obtained predictor variables that had a significant effect on the model were rice plant productivity, rice harvest area, community purchasing power index, percentage of households receiving food aid, and percentage of LBW babies. Ordinal probit regression modeling with the SMOTE approach has a classification accuracy of 74.51%.

Item Type: Thesis (Other)
Uncontrolled Keywords: Kawasan Timur Indonesia, ketahanan pangan, probit ordinal, regresi, SMOTE, Eastern Indonesia, food security, ordinal probit, regression, SMOTE
Subjects: H Social Sciences > HA Statistics > HA31.3 Regression. Correlation
H Social Sciences > HA Statistics > HA31.7 Estimation
S Agriculture > SB Plant culture > SB191.R5 Rice farming
Divisions: Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User: Muharinda Sugma Nur Ainin
Date Deposited: 08 Aug 2024 13:29
Last Modified: 08 Aug 2024 13:29
URI: http://repository.its.ac.id/id/eprint/114976

Actions (login required)

View Item View Item