关联规则挖掘算法的分析及应用-计算机应用技术专业论文.docxVIP

摘要 移动企业是典型的数据密集行业，拥有海量的用户数据。随着移动服务普及 程度和需求程度的日益提高，移动企业的竞争也日趋激烈。企业在竞争中获胜的 关键在于如何充分的利用移动数据库中的历史数据，挖掘出隐藏在数据背后对企 业有价值有意义的知识，帮助企业制定精准化营销策略，使得企业在竞争中处于 优势地位。 本文首先研究了数据挖掘的相关技术、方法和数据挖掘的挖掘过程，着重对 关联规则经典算法 Apriori 算法进行了深入研究，针对 Apriori 算法的缺点和不足， 它需要频繁的扫描数据库和生成大量的候选项集，提出了一种有效的改进算法。 该算法通过减少行扫描数和候选项集的生成数量，从而大大提高算法的效率，实 验证明该算法的执行效率明显高于经典 Apriori 算法。 然后，针对传统数量型关联规则划分边界过硬等问题，本文基于模糊集理论 及其性质对边界进行软化，应用模糊 c-均值聚类算法把原始数值型属性转化为相 应隶属度值，有效的避免尖锐边界问题。并通过设定相应阈值，减少支持度的计 算量，有效的减少模糊关联规则挖掘的时间，提高模糊关联规则挖掘效率。 最后，结合青海移动的海量历史数据，依据数据挖掘 CRISP-DM 流程模型， 将改进的布尔型关联规则算法和模糊关联规则算法应用到青海移动 GPRS 业务数 据中，从而挖掘出客户特征和移动 GPRS 业务之间的关联关系。挖掘结果证明了 算法的可行性与实用性，并且对移动的营销和策划人员具有积极的指导作用。 关键词：关联规则 Apriori 算法 模糊集 移动业务 Abstract Mobile Enterprises are typical data-intensive industries, which have mass user data. With the increased popularization and requirement of mobile services, the competitions of mobile enterprises are becoming fiercer. Meanwhile, the key role to excel in competitions is how to utilize the historical data in mobile database to mine the hidden knowledge, which are valuable and meaningful to help enterprises make accurate marketing strategy and consolidate their dominance. In this paper, we focus on the following contributions: First, we study the relevant techniques, methods and the procedure of data mining, especially for the classic Apriori algorithm. It needs to scan databases frequently and then generate a large amount of candidate itemsets. Aiming at the shortcomings and insufficiency of Apriori algorithm, an improvement algorithm is proposed to optimize it. The efficiency of the algorithm can be improved by reducing the times of scanning and controlling the generated candidate itemsets. The analysis and simulation results show the efficiency of the new algorithm is significantly higher than the classic Apriori algorithm. Second, because the traditional quantitative association rules have the problems of boundary division hard, this paper based on fuzzy set theory and their properties to soft the border and adop