Patterns from crime data using candidate generation approach

The rule mining algorithm generates rules from frequent patterns which are mined from Theft Crime dataset. The various combinations of the rules which is produced by the rule mining algorithm may be efficient or inefficient. It is waste of time to run all the rules. So, in order to validate the most efficient rule, the proposed algorithm applied the existing support and confidence measures and additionally one more measure information gain to add more values to the rule generation and validation. The proposed Theft Pattern Mining algorithm is adapted with the Improved Rule Mining algorithm and it is applied to the Tamil Nadu Theft Crime dataset.