Computing the probability on socio economic factors to predict the crime locations by means of joint probability based AMABC-FCIL

Abstract

The Frequent Itemsets Mining (FIM) is a demanding task common to several important data mining applications that look for interesting patterns within the databases. Several techniques have been proposed to mine the frequent closed itemsets. In this paper, we have proposed a frequent closed itemset mining technique based on probability. The socio economic factors are clustered with the help of the Adaptive Mutation based Artificial Bee Colony (AMABC) Algorithm after fetching them from the database. After clustering the attributes, the rules are generated and the Joint Probability Function (JPF) is computed. The rules satisfy the joint probability cutoff which is selected to construct the Frequent Closed Itemset Lattice (FCIL). The rules which satisfy the support threshold after constructing the FCIL are selected as the frequent closed itemsets. Finally, a testing process is included in which the known test data are provided. To analyze the performance of the proposed technique, certain performance metrics like time, memory, accuracy, lift and confidence rates are utilized and the performance of the proposed technique is improved with the existing Sliding with Itemsets Factor (SIF) based FCIL.