Analisis Migrasi dengan Model Loglinear dan Iterative Proportional Fitting

Abstract

Migration is one of demographic components that affect the population dynamics in a region. Data of migration flow can be presented in a contingency table. This paper discusses the application of loglinear model to describe the spatial structure of migration. The analysis of spatial structure of migration requires data on migration flows from a region of origin to its destinations. If the migration data are complete, then the parameters can be obtained directly. But, on the other hand, if the data are incomplete, then the parameters will have to be obtained by following steps. Firstly, assign the complete data as the standard value. Secondly, analyze data with the corresponding loglinear model. Finally, predict the incomplete values using iterative proportional fitting. A case study is conducted on migration flow data of four regions. Prediction of incomplete migration flows using iterative proportional fitting shows that the result is close to the actual data. Parameters of multiplicative loglinear model with cornered effect indicate that the first region is the region that has the greatest odds ratio. Whereas, parameters of multiplicative loglinear model with mean effect indicate that the third region is the most important region as origin and destination.