模糊关联规则在推荐系统的应用研究

发布时间：2019-01-25 20:25

【摘要】：伴随着信息产业化的发展,人们接受的信息越来越多了,这些信息已经超过了个人和系统所能接受和理解的范围,以致于影响到了我们的生活、学习、工作以及人际关系等。这就是信息过载,以及信息过载为我们带来的困扰。需求驱动着技术的发展,推荐系统就是这样产生了,它是为用户主动定位和推送其感兴趣的内容。个性化推荐系统广泛的应用到我们的生活、学习以及工业生产中,不但影响着我们的生活、学习和工作的方式,而且个性化系统广泛的应用促进了经济的发展,所以研究个性化推荐系统有实用意义。关联规则是先从历史数据中挖掘出来所有的频繁项集,其次再从频繁项集找出来所有的强关联规则。如今,关联规则已经成为了一种常用的推荐系统的方法。但是基于布尔型的关联规则只能处理布尔型的数据,不能处理数量型的数据,那么算法就有了一定的局限性,所以我们引入了模糊的概念。将模糊概念引入关联规则算法,应用到推荐系统中,提高推荐结果的合理性。同时关联规则有一个弊端,采用过大的频繁候选项集,这影响算法的效率。所以,本文提出来了基于决策树的模糊关联规则,并它应用到推荐系统中,这样不但提升了算法的效率,而且使推荐系统具备个性化和人性化。本课题以推荐系统为应用点,以关联规则为核心,以模糊理论和决策树为要点,主要介绍了关联规则算法、模糊关关联规则算法以及决策树算法的理论知识以及它们在推荐系统中的应用。研究成果有如下几个方面:1)研究关联规则与模糊关联规则的基本原理,介绍了apriori算法挖掘频繁集的过程,以及介绍了由频繁挖掘强关联的规则的过程。2)研究决策树算法的基本原理,并介绍了使用ID3算法和C4.5算法构建决策树的过程,并介绍了决策树的截枝和决策操作。3)研究基于决策树的模糊关联规则的算法,并着重介绍了决策树如何优化apriori算法的性能。4)将模糊关联规则和基于决策树的模糊关联规则应用到推荐系统中,并编程实现,在实际的应用中给出来两种模型的差异。
[Abstract]:With the development of information industrialization, more and more information has been accepted by people, which has exceeded the range of personal and system acceptance and understanding, which has affected our life, study, work and interpersonal relationships. This is information overload, and information overload for us to bring trouble. Requirements drive the development of technology, recommendation system is produced, it is active for users to locate and push the content of their interest. Personalized recommendation system is widely used in our life, learning and industrial production. It not only affects the way we live, study and work, but also promotes the development of economy. Therefore, the study of personalized recommendation system has practical significance. Association rules first mine all frequent itemsets from historical data, and then find all strong association rules from frequent itemsets. Nowadays, association rules have become a common method of recommendation system. But the association rules based on Boolean type can only deal with Boolean data, but not quantitative data, so the algorithm has some limitations, so we introduce the concept of fuzzy. The fuzzy concept is introduced into the association rule algorithm and applied to the recommendation system to improve the rationality of the recommendation results. At the same time, there is a disadvantage of association rules, which affects the efficiency of the algorithm by using too large set of frequent candidate items. Therefore, a fuzzy association rule based on decision tree is proposed in this paper, and it is applied to the recommendation system, which not only improves the efficiency of the algorithm, but also makes the recommendation system personalized and humanized. Based on recommendation system, association rules, fuzzy theory and decision tree, this paper mainly introduces the algorithm of association rules. The theoretical knowledge of fuzzy association rules algorithm and decision tree algorithm and their application in recommendation system. The research results are as follows: 1) the basic principles of association rules and fuzzy association rules are studied, and the process of mining frequent sets by apriori algorithm is introduced. This paper also introduces the process of mining strong association rules frequently. 2) the basic principle of decision tree algorithm is studied, and the process of constructing decision tree using ID3 algorithm and C4.5 algorithm is introduced. It also introduces the pruning and decision operation of decision tree. 3) the algorithm of fuzzy association rules based on decision tree is studied. How to optimize the performance of apriori algorithm based on decision tree is introduced. 4) the fuzzy association rules and fuzzy association rules based on decision tree are applied to the recommendation system and implemented by programming. The differences between the two models are given in the practical application.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.13;TP391.3

【相似文献】