Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/114125
Title: Identifikasi Faktor yang Memengaruhi Anak Putus Sekolah di Provinsi Daerah Istimewa Yogyakarta Menggunakan Regresi Logistik Biner
Authors: Anisa, Rahma
Alamudi, Aam
Hasanah, Malikhatul
Issue Date: 2022
Publisher: IPB University
Abstract: Anak putus sekolah merupakan anak yang tidak mampu menyelesaikan pendidikannya dan berhenti sekolah sebelum masa belajarnya berakhir sehingga belum memiliki tanda tamat belajar atau ijazah pada jenjang pendidikan tersebut. Provinsi DIY merupakan salah satu provinsi yang memiliki kasus putus sekolah dan mengalami peningkatan dari tahun 2018 sampai 2020. Untuk itu, penelitian ini bertujuan memodelkan kejadian anak putus sekolah di Provinsi DIY menggunakan regresi logistik biner dengan menerapkan metode SMOTE untuk menangani ketidakseimbangan data sehingga diketahui faktor-faktor yang memengaruhinya. Regresi logistik biner digunakan karena peubah respon bersifat biner (Y = 1 jika putus sekolah dan Y = 0 jika masih aktif bersekolah). Data penelitian menggunakan data sekunder hasil Survei Sosisal Ekonomi Nasional 2021 yang dilakukan oleh Badan Pusat Statistik wilayah Provinsi DIY. Hasil penelitian menunjukkan faktor yang berpengaruh, yaitu pendidikan kepala rumah tangga, rata-rata pengeluaran rumah tangga perkapita, status bekerja anak, dan status penerima PIP dengan ketepatan klasifikasi model berikut : akurasi 79,16%, F1-score 17,19%, dan AUC 83,22%.
Dropout schools are children who are unable to complete their education and leave school before the end of their study period so that they do not obtain certificate from school or a diploma at that level of education. DIY Province is one of the provinces that has dropout cases increasing from 2018 to 2020. For this reason, this study aims to model the incidence of school dropouts in DIY Province using binary logistic regression by applying the SMOTE method to handle data imbalances so that the factors that influence it are known. Binary logistic regression was used because the response variables were binary (Y = 1 if they dropped out of school and Y = 0 if they were still in school). This research used secondary data from the 2021 National Socio-Economic Survey conducted by the Central Statistics Agency for the DIY Province. The results showed that the influencing factors were the education of the head of the household, the average household expenditure per capita, the child’s working status, and the status of PIP recipients with 79,16% accuracy, 17,19% F1-score, and 83,22% AUC.
URI: http://repository.ipb.ac.id/handle/123456789/114125
Appears in Collections:UT - Statistics and Data Sciences

Files in This Item:
File Description SizeFormat 
Cover.pdf
  Restricted Access
Cover206.75 kBAdobe PDFView/Open
G14180018_Malikhatul Hasanah.pdf
  Restricted Access
Fullteks2.19 MBAdobe PDFView/Open
Lampiran.pdf
  Restricted Access
Lampiran765.79 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.