基于关联规则的蚁群算法在交叉销售中的应用研究

基于关联规则的数据挖掘技术对交叉销售的策略制定有着重要作用.针对关联规则算法需多次扫描数据库和可能产生庞大的候选集等问题,提出了一种新的基于关联规则的蚁群算法,利用蚁群算法中的信息素因子得到最强关联规则,从而找到具有商业价值的最大频繁项集.实验结果表明新算法充分发挥了蚁群算法自组织,多样性,并行性等优点,不仅增加了频繁项集的生成数量,而且较大的提高了它的生成效率.

Application Research of Ant Colony Algorithm Based on Association Rules in Cross-selling

Data mining technology based on association rules for cross-selling plays an important role in strategic plan- ning. Because association rules algorithm need to scan the database for multiple times and may produce large candidate sets and other issues, a new ant colony algorithm based on association rules is proposed, and the pheromone factor of ant colony algorithm is used to get the strongest association rules, thus the maximal frequent itemsets for commercial value are found. Experimental results show that the new algorithm fully realizes the features of self-organization, diversity, parallelism of the ant colony algorithm, etc,which not only increases the number of frequent itemsets, but also greatly improves the efficiency of its generation.