Mining Optimized Association Rules For WEB Page Based On Genetic Algorithm

In the era of rapid development of information technology,people browse web pages to learn knowledge or understand some areas of interest.Web pages have also become an important factor to reflect user interest and build web page framework.Therefore,mining the association between web pages is of great significance to reconstructing a web page framework.Quantified browsing frequency reflects various degree of users' interest.However,quantified values can not be a good measure of user's interest.In this paper,quantified browsing frequency is described as the corresponding fuzzy language variable.And an optimized association rule mining algorithm based on genetic algorithm is proposed.First,the browsing frequency is transformed into a fuzzy set by using a random membership functions,but the random membership functions has a great influence on the extraction of rules.Therefore,the genetic algorithm is used to optimize the membership functions in this paper.The membership functions are encoded into the corresponding chromosomes in genetic algorithm,and an optimal chromosome is found by the cross-mutation operation.This optimal chromosome can better describe the user's browsing frequency and better reflect the user's interest level.Secondly,the user browsing frequency is transformed into a corresponding fuzzy set according to the optimal membership function.The confidence levels of all possible association rules are calculated,association rules with high confidence level are retained.In the traditional way,the association rules are expressed as follows: BA ?.While fuzzy association rules in this paper are represented as:)()(21BA?? ?,where 1?? 2? denote fuzzy linguistic variables,indicating the lowhigh-degree of users' interest.In contrast,fuzzy association rules in this paper are able to depict the degree of users' interest on web pages preferably.