Lumbung Pustaka UNY: No conditions. Results ordered -Date Deposited. 2024-03-28T11:42:49ZEPrintshttp://eprints.uny.ac.id/apw_template/images/sitelogo.pnghttps://eprints.uny.ac.id/2015-07-15T21:40:30Z2015-07-15T21:40:30Zhttp://eprints.uny.ac.id/id/eprint/23641This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/236412015-07-15T21:40:30ZAUTOLOGISTICS MODELS FOR MODELLING AND MAPPING INFANT MORTALITY IN INDONESIAAccording to Indonesia Demographic and Health Survey (IDHS) data, the highest child mortality occurred during the first year of age of infant. Infant mortality is an important indicator that must to be monitored seriously. The mortality is associated with several determinants, such as the infant’s characteristics, maternal and fertility factors, housing condition, and also geographical area. The aim of this research is to develop models that can be used to explain the effects of explanatory variables on infant mortality in Indonesia. It is also the aim of this research to develop a thematic map describing the distribution pattern of infant mortality probabilities at district level across the country. The response variable is a binary categorical variable with two outcomes, success and failure. The outcome is success if the infant died before achieving one year of age, and failed if the infant is still alive after one year of age. Modeling is using Logistic Regression model and Autologistics Regression model. The results showed that the Autologistics Regression model fitted the data reasonably well, all of the explanatory variable affect infant mortality, except infant’s sex. The results also showed that the probabiity of infant mortality was higher in Kalimantan island and Papua island.
Keyword : Infant, mortality, autologistic, binary, IDHS. Ray Sastri. Khairil Anwar Notodiputro. Indahwati2015-07-15T21:29:16Z2015-07-15T21:29:16Zhttp://eprints.uny.ac.id/id/eprint/23637This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/236372015-07-15T21:29:16ZMODELLING THE AVERAGE SCORES OF NATIONAL EXAMINATION IN WEST JAVAFormal education in Indonesia is commonly divided into stages such as preschool, primary school (SD), Secondary School (SMP-SMA), and universities/colleges. Indonesian government has been taking serious efforts on how to improve the quality of education in Indonesia. The roadmap for continous improvement of education quality can be designed based on the results of National Examination (UN) taken regularly by high school students.
This research was aimed at exploring informations on how the scores of UN can be linked with other explanatory variables. A panel data which consists of average scores of UN for all public senior high schools (SMA Negeri) in West Java Provinces during 2011-2013 and other related variables such as total scores of accreditation, regional domestic product, human development index, scores of school’s facilities and its infrastructure, scores of school’s educators, average scores of final school exams, were used in this research. The average scores of UN in this case were dependent on variations between high schools and time periods as well as other explanatory variables in which the effects were either fixed or random. The data of this research was modelled with linear mixed models and using the Generalized Estimating Equation (GEE) approach. Both linear mixed models and GEE have been commonly used to analyse the panel data.
This paper showed that the GEE provided a model of better performance than the linear mixed models in explaining the variability of the response variable which was the average scores of UN. The GEE also showed significant correlation between explanatory variables and the response.
Key words: fixed effects, GEE, linear mixed model, national examination, random effects.. Karin Amelia Safitri. Khairil Anwar Notodiputro. Anang Kurnia2015-07-15T12:31:09Z2015-07-15T12:31:09Zhttp://eprints.uny.ac.id/id/eprint/23630This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/236302015-07-15T12:31:09ZMODELLING CASES OF LOW BIRTH-WEIGHT INFANTS WITH GENERALIZED LINEAR MIXED MODELLow Birth-Weight (LBW) is defined as a birth weight of a live-born infant of less than 2.500 grams regardless of gestational age. The causes of LBW cases can be grouped into two main causes: premature birth and case of small for gestational age (SGA). There are many risk factors that can induce directly or indirectly so that these causes may occur. Case of LBW is associated with infant mortality, infant morbidity, inhibited growth and slow cognitive development, also chronic diseases in later life.
To suppress rate of LBW first we must estimate the rate correctly. Data of LBW comes from Indonesian Health and Demographic Survey (IDHS) 2012 which is divided into 3 groups: written (measured accurately), recall (measured inaccurately) and not weighed (not measured). Published national rate of LBW is 7.3% with provincial rates fall between 4.7-15.7 %. The estimation came from only 2 former groups without consideration of assumed difference accuracy on second group.
To estimate the difference and the rate of the third group, Generalized Linear Mixed Model (GLMM) is used with live-born infants as observation units because observations from the same sampling unit tends to correlate due to multistage sampling design.
The result of the model at α = 0.05 is highly-significant, with fixed effect variables that are statistically significant to the case of LBW are Estimated Size, Preceding Interval, Pregnancy Complication, Mother’s Age, Province and Education. Higher portion of variance component is on the G-side as a result of multistage sampling, with Household level has highest within variance. On the R-side, recall group data has higher variance than written group. It is an indication of lower accuracy of the birth weight data on this group. Based on the model, estimation of LBW rate including not weighed group result 7.96% slightly higher than direct estimate.
Keywords: Low Birth-Weight, GLMM, Logistic Regression, IDHS 2012. Antonius Benny Setyawan. Khairil Anwar Notodiputro. Indahwati2015-07-10T04:31:53Z2015-07-10T04:31:53Zhttp://eprints.uny.ac.id/id/eprint/23300This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/233002015-07-10T04:31:53ZCOMPARISON WILLIAMS METHOD AND BETA-BINOMIAL IN OVERDISPERSION OF LOGISTIC REGRESSION: A CASE OF INDONESIA GENERAL ELECTION DATA 2014Democratization in Indonesia so far has resulted in increasingly rational voters. The rational voters in each district or city of Indonesia are varied due to many factors. The system of election in Indonesia today is direct election system in which every citizen has freedom to vote the preferred candidates or even not to vote at all. There were 12 political parties participated in the legislative election in 2014, whereas in the presidential election there were two pairs of president and vice-president candidates competed.
This research was aimed to obtain models, at the district level, that properly relate the votes were gained by the two candidates and other variables such as human development index, the results of legislative election as specially coalition of political parties voting results. Since the vote data was binary and showed over-dispersion then a logistics model accounting for over-dispersion was utilized. An over-dispersion problem is present whenever observations which might be expected to correspond to the binomial distribution may have greater variance than ni πi (1-πi).In this research the William’s method and beta-binomial regression were used to overcome the problem. The result showed that the Williams method provided better estimates when was compared to beta-binomial regression.
keyword: Logistic regression, Overdispersion, Williams’Method, Beta-Binomial Regression, General Election. Firman Hidayat. Khairil Anwar Notodiputro. Bagus Sartono