Optimizing mining association rules based on Artificial Neural Network

The basic function of a neural system is to intelligent learning from specific examples known as neurons. It has great pattern adaptive capability that may be used to judge between old model and well model. Neural systems have many characteristics such as autonomous, uniqueness, recognition of foreigners, noise tolerance, and distributed detection. Inspired by neural network system, Artificial Neural Network has emerged during the last decade. It is incited by many researchers to build, study, and design neural-based models for a variety of application regions. Artificial neural system can be defined as adaptive model that is inspired by neural network, observed neural functions, mechanisms and principles. Association rule mining is one of well researched and the most important techniques of datum mining. The purpose of association rules is to refine interesting correlations, associations, frequent patterns, or casual constructions in sets of aims in other datum repositories or the transaction databases. Association rule is widely used in various regions such as telecommunication network, inventory control, intelligent decision, risk management and market analysis etc. Artificial Neural Network is the most widely used algorithm for mining the association rules. In this paper, Artificial Neural Network is studied and optimized based on classification system. The performance of the ANN based on classification system is evaluated by varying number of generations and computing accuracy at different factors. Three standard datum had been used to computer the accuracy. The test result shows that the system can give highest accuracy more than o.4.