多模式下基于频繁关联的电子易货匹配算法

发布时间：2018-01-28 00:45

本文关键词： 易货电子易货电子商务频繁关联　出处：《四川师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网的兴起,电子易货得到了迅猛的发展,通过电子易货的交易模式可以使得国际和企业之间的资源得到再次交换和匹配,能够极大的促进国家的经济发展和企业之间的业务流动。电子易货交易中最大的难点是将易货者之间的供需进行匹配,这需要通过合理有效的易货匹配算法进行寻找。本文针对电子易货资源匹配问题,将一次交易涉及交易数量最多作为算法的总目标,研究了一整套以供求关系有向图模型为基础的电子易货交易平台交易数据的挖掘与匹配算法。提出一种全新的基于频繁关联的电子易货模型,并在单次交易匹配匹配和多人协同匹配两种模式下进行了应用。使用电子易货匹配算法能够有效的传统易货平台过分依赖易货经济人,商品无法自动匹配交易等问题。本文研究的工作主要包括以下两个方面:(一)对每个易货者每次只能提交一个需求和一个供给资源的单次交易匹配模式易货算法进行改进。本文提出基于频繁关联的易货算法,首先将电子易货中的冗余数据进行清理,排除不能成环的交易记录;其次,根据资源匹配中各个易货者所提供的资源与其他易货者资源关联的频数(Frequency)对参与的易货者进行排序,选取出评分最高的易货者作为遍历环的开头,并在随后的每一个分支节点时都根据排序的评分来选择相应的易货者作为优先遍历者,快速找出最长不相交环。通过实验与传统遍历算法相比取得较好的改进效果。(二)对多人协同易货模式的易货算法进行实现。由于涉及易货者之间的协同交易,其解决方案是一种NP难题,匹配实现的主要难度在于方案的最终结果不能确定是否为可行的。在本文中,首先,将电子易货中的冗余数据进行清理;其次,根据资源匹配中各个易货者所提供的资源与其他易货者资源关联的频数(Frequency)对参与的易货者进行排序,选择拟匹配队列的方式将易货者的需求逐步进行满足,从而全面搜索多人协同模式下的交易可能方案。实验证明该算法能够在多人协同模式的数据中得到良好的匹配结果。
[Abstract]:With the rise of the Internet, electronic barter has been a rapid development, through the electronic barter transaction model can make international and enterprise resources to be re-exchanged and matched. Can greatly promote the economic development of the country and business flows between enterprises. The biggest difficulty in electronic barter transactions is to match the supply and demand between barter. This needs to be found through a reasonable and effective barter matching algorithm. Aiming at the problem of electronic barter resource matching, the maximum number of transactions involved in a transaction is regarded as the total target of the algorithm in this paper. This paper studies a set of data mining and matching algorithms of electronic barter trading platform based on the digraph model of supply and demand relationship, and proposes a new electronic barter model based on frequent correlation. It is applied in the two modes of single transaction matching and multi-person co-matching. Using electronic barter matching algorithm, the traditional barter platform can rely too much on barter. The work of this paper mainly includes the following two aspects: (1). This paper improves the single transaction matching pattern barter algorithm that each barter can only submit one demand and one supply resource at a time. This paper proposes a barter algorithm based on frequent association. First of all, the redundant data in electronic barter is cleaned out to eliminate the transaction records that can not be looped. Secondly, according to the resources provided by each barter in the resource matching and other barter resources associated with the frequency of frequent) to participate in the barter ranking. The barter with the highest score is selected as the beginning of the traversal loop, and the corresponding barter is selected as the priority traversal at each subsequent branch node according to the ranking score. Find out the longest disjoint loop quickly. Compared with the traditional ergodic algorithm, the experiment has better effect. (2). This paper implements the barter algorithm of multi-person cooperative barter pattern, which involves the cooperative transaction among barter. The solution is a kind of NP problem, the main difficulty of matching implementation is that the final result of the scheme can not be determined whether it is feasible or not. In this paper, firstly, the redundant data in electronic barter are cleaned up. Secondly, according to the resources provided by each barter in the resource matching and other barter resources associated with the frequency of frequent) to participate in the barter ranking. Choosing the way to match the queue will gradually satisfy the barter's needs. The experiment results show that the algorithm can get a good matching result in the data of multi-person cooperative pattern.
【学位授予单位】：四川师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP301.6;F713.36

【参考文献】