Generation of Negative and Positive Association Rules Using Modified Algorithm

The primary objective of this work is the optimization of association rules by modified genetic algorithm for better rule prediction. The proposed algorithm is an extension to genetic algorithm which provides modified genetic algorithm-based optimizer. Modified genetic algorithm is applied to find out optimal rules that use common genetic algorithm operators to find a population of solutions, based on the fitness function value. The modified genetic algorithm-based optimizer selects the most useful set of rules with comparatively less number of iterations required for standard genetic algorithm. This improved technique efficiently discovers all correct association rules, removing a major bottleneck from earlier mining approaches. These optimized rules produce effective results in making any decision in support system. This work also introduces the basics of rule generation in data mining.