Recognizing poverty pattern in Central Java using Biclustering Analysis

Poverty is a complex and multidimensional problem and becoming a development priority. In analyzing the pattern of poverty in a region, one of the statistical procedures that is usually used is the cluster analysis. However, it does not consider the different levels of performance by region in different characteristics at a particular time. In this study, an alternative approach, namely Cheng and Church’s biclustering algorithm, was used to simultaneously identify the poverty pattern in Central Java by region and poverty dimension variables. Using this algorithm, we found two biclusters with different characteristics. The first bicluster represents the general condition of poverty in Central Java, but they are better in labor and housing quality indicators. While the second bicluster is poorer in some indicators of labor, education, and housing quality than Central Java.