Crime Identification using FP-Growth and Multi Objective Particle Swarm Optimization

摘要

Crime detection problem can be modeled by using data mining techniques. Crimes are the community trouble for which our society has to pay in several ways. About 10 percent of the criminals commit about 50% of the crime. Here we are providing a help to criminal analysts to detect the crime in terms of crime pattern. Crime experts can be helped by discovering new patterns on the basis of crime type and location where the crimes occur. For drawing out crime pattern, pattern based approaches can be used. In this paper, optimized patterns are generated by using multi objective particle swarm optimization. MOPSO optimized the patterns generated by the working of FP-Growth algorithm on criminal database. The working of FP-Growth and MOPSO algorithm on crime data is also explained by flowchart and example. Empirical results are also presented in order to direct future audience about upcoming decisions or research concerning this model. Results indicate that our proposed algorithm is more promising and will always make sure the consumption of a reasonable percentage of the crime dataset, in terms of execution time and useful, non-redundant patterns.