ESTIMASI PARAMETER REGRESI LOGISTIK BINER DENGAN MAXIMUM LIKELIHOOD ESTIMATION DAN GENERALIZED METHOD OF MOMENT (Studi Kasus: Faktor-faktor yang mempengaruhi kemiskinan di Provinsi Jawa Timur tahun 2018)

ANJAR MEI ANGGORO WATI, NIM. 17106010034 (2021) ESTIMASI PARAMETER REGRESI LOGISTIK BINER DENGAN MAXIMUM LIKELIHOOD ESTIMATION DAN GENERALIZED METHOD OF MOMENT (Studi Kasus: Faktor-faktor yang mempengaruhi kemiskinan di Provinsi Jawa Timur tahun 2018). Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (ESTIMASI PARAMETER REGRESI LOGISTIK BINER DENGAN MAXIMUM LIKELIHOOD ESTIMATION DAN GENERALIZED METHOD OF MOMENT (Studi Kasus: Faktor-faktor yang mempengaruhi kemiskinan di Provinsi Jawa Timur tahun 2018)) 17106010034_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (2MB) | Preview
	Text (ESTIMASI PARAMETER REGRESI LOGISTIK BINER DENGAN MAXIMUM LIKELIHOOD ESTIMATION DAN GENERALIZED METHOD OF MOMENT (Studi Kasus: Faktor-faktor yang mempengaruhi kemiskinan di Provinsi Jawa Timur tahun 2018)) 17106010034_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf - Published Version Restricted to Registered users only Download (4MB)

Abstract

Binary logistic regression is logistic regression where the dependent variable is dichotomous data consisting of two categories. To determine the binary logistic regression model, an estimated parameter is needed to obtain information about a population. The estimation of binary logistic regression parameters used is the Maximum Likelihood Estimation (MLE) method with Newton Raphson iteration and Generalized Method of Moment (GMM) with Iterative Reweighted Least Square. This research takes a case study of poverty data in East Java Province in 2018 which consists of 38 districts/cities, with the dependent variable in the form of binary data with poor and not poor categories. The estimation results obtained from MLE and GMM methods produce a number of different significant variables. From the MLE method, there is one significant independent variable included in the model, namely the household receiving PKH variable. Meanwhile, from the GMM method, there are two significant variables included in the model, namely household that has a computer variable and the household receiving PKH variable. Judging from the value of the classification accuracy of the MLE method of 81.57% and the accuracy of the classification of the GMM method of 73.68%, so that the estimation of binary logistic regression parameters in the case study of factors that affect proverty in East Java Provience in 2018 using the MLE method is better than using the GMM method.

Share this knowledge with your friends :

Actions (login required)

View Item