PEMODELAN REGRESI ROBUST M-ESTIMATOR DALAM MENANGANI PENCILAN (STUDI KASUS PEMODELAN JUMLAH KEMATIAN IBU NIFAS DI JAWA TENGAH

Alan Prahutama(1*), Agus Rusgiyono(2), Dwi Ispriyanti(3), Tiani Wahyu Utami(4)


(1) 
(2) 
(3) 
(4) 
(*) Corresponding Author

Abstract


Regression analysis is statistical method that used to model between predictor variables and response variables. In the regression model, the residual assumed normal distribution, non-autocorrelation, and homoscedasticity. When the assumptions doesn’t fulfilled, then the measurement of goodness not well enough. One of the causes may be outlier of data. Handling the outlier can be used robust regression, which one of method is robust M-estimator.   In this article, we purposed modelling the number of maternal postpartum in Central Java province with predictor variables are the percentage of pregnant who consumed Fe tablet (X1), the percentage of household whom applied clean and health lifestyle(X2), and the percentage of pregnant who First visited to midwife of doctor (K1) (X3).  In the multiple regression only X3 was significantly with R-square was 14.25209%, and Mean Square Error (MSE) was 20.4177. Moreover, in outlier detection, there were two outlier in the data, then modelled with Robust M-estimator. The measurement of goodness used R-square of regression robust M-estimator was 21.74% with MSE was 15.02766. Robust M-estimator regression resulted better model than multiple regression to model the number of maternal postpartum in Central Java Province.

Full Text:

PDF

Article Metrics

Abstract view : 86 times
PDF - 29 times

DOI: https://doi.org/10.26714/jsunimus.9.1.2021.35-39

Refbacks

  • There are currently no refbacks.


Copyright (c) 2021 Jurnal Statistika Universitas Muhammadiyah Semarang

Editorial Office:
Department of Statistics
Faculty Of Mathematics And Natural Sciences 
Universitas Muhammadiyah Semarang

Jl. Kedungmundu No. 18 Semarang Indonesia

Published by: 
Department of Statistics Universitas Muhammadiyah Semarang

View My Stats

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.