Lumbung Pustaka UNY: No conditions. Results ordered -Date Deposited. 2024-03-28T16:13:22ZEPrintshttp://eprints.uny.ac.id/apw_template/images/sitelogo.pnghttps://eprints.uny.ac.id/2015-07-15T21:38:04Z2015-07-15T21:38:04Zhttp://eprints.uny.ac.id/id/eprint/23640This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/236402015-07-15T21:38:04ZCLUSTERING DISTRICTS BASED ON INFLUENCED FACTORS OF CHILD UNDERNUTRITION (STUNTING)One indicator of undernutrition problems is stunting. In Indonesia, the prevalence of stunting of under 5 years children has increased from 35.6% (2010) to 37.2% (2013) and there are differences among provinces. These disparities are likely to be greater when the level is seen in smaller areas such as inter regencies/cities. The purpose of this study. is to group the regencies/cities based on stunting factors.
Prevalence factors of stunting of children under 5 years have different measurement scale (numerical and categorical). Two step cluster method is a method that is designed to handle a large number of objects, especially on the objects that have the problem of continuous and categorical variables. This method also assumes independent variables involved.
Based on BIC values, it can be concluded that the optimal number of cluster is 2, with cluster quality is fair. The first clusters consist of 223 counties and cities, while the second comprises 272 ones. In addition, there are two outliers, District Ndunga and District Lanny Jaya located in the Province of Papua. The first cluster has average value greater than the national average in the socioeconomics and child diet, in the other hand, the second cluster has average value greater than the national average in mother enviromental and health care, and the outliers has an average above the national average for all variables except on expenditure per capita, the average member of the household, and calorie consumption.
Keywords: clustering, malnutrition, stunting, two step cluster method. Nurul Istiqomah. Hari Wijayanto. Farit M. Afendi2015-07-15T12:42:07Z2015-07-15T12:42:07Zhttp://eprints.uny.ac.id/id/eprint/23633This item is in the repository with the URL: http://eprints.uny.ac.id/id/eprint/236332015-07-15T12:42:07ZMULTIVARIATE ADAPTIVE REGRESSION SPLINE (MARS) FOR MODELLING OF CHILD LABOR IN JAKARTAA Child is defined as male or female under the age of 18 years unless under the law of maturity has been reached earlier. Based on data from the ILO, there were four million child labors in Indonesia in 2002. Jakarta as the capital city of Indonesia has a high economic growth in 2010 by reaching 6.51 percent (LKPJ 2010). Behind that growth, Jakarta has a problem of high number of child workers as well, which is about 93 571 children in 2010.
Some researchers often use regression analysis to determine the description of the factors that contribute to a response variable. Regression analysis has several assumptions that must be met, while research in the social subjects often violates those assumptions. To overcome this limitation required nonparametric method that is not tied to the assumption. One method is non-parametric regression Multivariate Adaptive Regression Spline (MARS). MARS method is an approach for nonparametric regression model that can accommodate multicollinearity in the model.
This study uses secondary data drawn from SUSENAS in 2013 in DKI Jakarta. Response variable used is the status of work in children aged 10-17 years, while the predictor variables are fifteen variables that represent the characteristics of children and household. Based on the results of processing with MARS, obtained models are affected by Status of Child’s Education, Child’s Education, Child Live with Parent, Education of Head of Household, and etc.
Key words: child labor, MARS, SUSENAS, Jakarta. Dimas Adiangga. Hari Wijayanto. Bagus Sartono