基于 FP－growth算法的数据挖掘实例研究

摘要

Association rule mining algorithm is applied to discovering interesting links which are hidden in the massive amounts of data and is widely used in the fields of Finance,Biology,Business,Medicine,etc.FP-growth algorithm is one of the classical algorithms which just scan the database for two times.At the same time,it has the attribute of compressing the size of the search data sets,so it is suitable for rules of different lengths.In this paper, the procedure of FP-growth was analyzed and applied to processed music data.It dug out accurate data and rules which were in line with users'demands through their attribute information.Finally it recommended music and the ad to users.Experiment results show that the running time of FP-growth is one order lower than Apriori.In addition,it is feasible to apply FP-growth to the music recommendation.%关联规则挖掘算法致力于发现隐藏在海量数据中的有趣联系，被广泛应用在金融、生物、商业、医学等领域，FP－growth算法是其中的经典算法，只需扫描两遍数据库，可以压缩被搜索数据集的大小，能够适应不同长度的规则。文中分析了FP－growth算法的执行步骤并将其应用于预处理的音乐数据，结合用户属性挖掘出符合用户需求的精准数据和规则并进行音乐推荐和广告推送。实验表明FP－growth算法的运行时间比Apriori算法大约少了一个数量级，且在音乐推荐方面具有可行性。