Applying Opposition Attributes Clustering to Matching Problems Using Bartering as an Example

• Main Authors: Jui-Yin Tai, 戴瑞瑩
• Other Authors: Jin-Ling Lin
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/95372307702643122255

碩士 === 世新大學 === 資訊管理學研究所(含碩專班) === 103 === This thesis focuses on how to efficiently find exchange cyclic patterns in the large database, which contains opposition attributes. First, "Opposition Attributes Clustering" is applied into the large database, with opposition attributes, to find appropriate clusters for matching. Then, perform "Exchange Cyclic Pattern Improving Search" on these clusters, found in opposition attributes clustering, to search exchange cyclic patterns. Combining opposition attributes clustering and exchange cyclic patterns improving search not only can find the exchange cyclic patterns rapidly but also can enhance the opportunity of matching To validate the performance and practicality, a computer program, based on the proposed opposition attributes clustering and exchange cyclic pattern improving search, was implemented. Simulated and experimental results shows the proposed approach could rapidly and efficiently find the exchange cyclic patterns. The proposed approach could also further applied into matching problems, such as the kidney exchange problems, the government of tax investigation problems, and the bartering problems.