The Applied Research Of Association Rules Mining Based On Colony Algorithm In Marketing

In the network, new products are developed, the market is increasinglycompetitive and fast-saturated in the background of modern marketing, e-commercecontinues to expand the size, number and types of commodities have been rapidgrowth, this process a lot of extraneous information for products will undoubtedlylose consumers in the overload information, so in the continuous loss of massive, anadvanced business intelligence platform based on data mining arise. We can use theplatform support aids. Whether network business or entity sales marketing enterprises,many began to combine the cross-selling and data mining techniques, by finding theoptimal combination of product sales and cross-selling target customers seeking toachieve a win-win for consumers and businesses, improve customer loyalty,andincrease corporate profits and market share.Association rules mining technology as a good solution to the above problem isone of data mining techniques and has been widely appreciated and studied. Thetraditional association rule mining algorithm needs to scan the database multiple times,so it spends a lot of execution costs, and thus how to reduce the number of scans ofthe database and how to improve the efficiency of the algorithm, still has a veryimportant significance.In this paper, ant colony algorithm has the advantages of this new algorithm, insome studies on the basis of existing, proposed ant colony algorithm based onassociation rule mining, specific studies as follows:(1)Between the different commodities (items) in the transaction database,constructed with the similar TSP solution space, put every item as each vertex of anundirected graph, each vertex set up the interest measure, impact the ants search pathselection, improving speed and quality of generated association rules.(2)Set different rule’s list to generate the positive and negative association rules,due to differences in the degree of interest measure arising from the classification isplaced in the rule set. In no weights to each side of the undirected graph is set, that is,the frequency between the vertices, as pheromone of ant colony algorithm, proposedmining association rules algorithm based on improved ant colony algorithm.(3)Use a collection of database product attributes and attribute classification,structure the solution space of ants traverse, classification rule mining is proposedbased on ant colony algorithm to dig merchandising business information needed from the perspective of the goods themselves attribute.(4)In this paper, put the sale of a music CD as example, through the experimentalresults and their analysis, validate the association rules and classification rule miningalgorithm based on ant colony algorithm for proposed algorithm, which is moreefficient than the past related algorithm whether in the quality of the rules generatedor efficiency.