AB Initio Analysis of Philippine Crime Statistics
Author: Jose Ramon G. Albert and Patrick Dennis L. Solosa
Here we provide a practical exposition of Bayesian methods and recent developments in computations for modeling crime statistics in the Philippines. The conceptual framework of Bayesian data analysis (vis a vis classical statistical analysis) is firstly discussed. Then contemporary Bayesian computing and modeling is illustrated through the use of directed acyclic graphs and the WinBUGS freeware. We provide a way of disaggregating the crime statistics for the Autonomous Region of Muslim Mindanao (ARMM) in the time periods before this region was formally established and of showing that most of the regional crime rates are on a decreasing trend.