REGRESSION ANALYSIS SPATIAL USING SIMULTANEOUS AUTOREGRESSION AND CONDITIONAL AUTOREGRESSION METHOD

• Main Author: Mulyarahardja Madjiah, Andreas
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39137
• Institution: Institut Teknologi Bandung

Spatial regression is a regression that takes into account spatial factors in the model. Ordinary regression cannot be done on data in the presence of spatial influences due to regression’s asumption of independence error which is violated. Spatial regression can be done on data lattice so it can be applied to see the relationship of fisheries to the percentage of poverty. Poverty itself is a serious problem that occurs in every country in the world. Poverty means someone's inability to fulfill their needs. One of the factors that influenced proverty is the quality of the workforce and inefficient use of human resources. Indonesia’s geographical location as one of the largest maritime countries that has a vast and rich sea makes it easy for the Indonesian population to make a living in the fisheries sector. Based on this, the influence of fisheries want to be seen in the percentage of proverty in Indonesia, especially the Java islands. The spatial regression model used is Simultaneous Autoregressive (SAR) and Conditional Autoregressive (CAR). Maximum Likelihood Estimation (MLE) is used to estimate the parameters in this model. The Moran’s I test was used to test for spatial factors in the response variable. The results obtained show that there is spatial factors in data with significant Moran’s I test result. The spatial regression model is better than ordinary regression with some fisheries having a significant influence on the percentage of proverty.