基于偏序的序列模式挖掘算法研究

发布时间：2018-04-18 08:25

本文选题：数据挖掘 + 序列模式　；参考：《四川师范大学》2017年硕士论文

【摘要】：序列模式挖掘是数据挖掘研究的一个重要领域,它在计算机入侵检测、WEB点击流分析、DNA序列分析和购物篮分析等领域有广泛应用。本文首先讨论了序列模式挖掘的基本理论,分析了基于频繁序列和频繁闭序列思想的几个经典序列模式挖掘算法。然后,系统讨论了偏序模式挖掘的基本理论,分析和研究了全局偏序和频繁闭偏序挖掘的典型算法及其优缺点。基于闭序列模式概念和TS-FCPO算法思想,论文提出了挖掘频繁等长闭偏序的EL-FCPO算法。它利用等长序列模式之间匹配时只有序列中的对应的位置才可能匹配的特点,改进了TS-FCPO算法寻找匹配位置时的枚举方式,以提高挖掘速度。等长闭偏序继承了闭偏序的所有优点,且可以表示出具有等长特点的闭序列之间的关系及其确定性趋势。由于等长偏序充分考虑了序列长度相等的序列之间只有序列上对应位置才有可能匹配的特点,使得算法在运用于等长偏序应用场景时的实际复杂度能够降低,时间效率有明显提升。论文成果可以应用于制定图书在线组合促销方案、旅游线路规划和攻略等方面。
[Abstract]:Sequential pattern mining is an important research field in data mining. It is widely used in computer intrusion detection Web click-stream analysis DNA sequence analysis and shopping basket analysis.In this paper, the basic theory of sequential pattern mining is discussed, and several classical sequential pattern mining algorithms based on frequent sequence and frequent closed sequence are analyzed.Then, the basic theory of partial order pattern mining is discussed systematically, and the typical algorithms of global partial order mining and frequent closed partial order mining are analyzed and studied as well as their advantages and disadvantages.Based on the concept of closed sequence pattern and the idea of TS-FCPO algorithm, this paper proposes a EL-FCPO algorithm for mining frequent equal-length closed partial order.It makes use of the feature that only the corresponding position in the sequence can be matched when matching the pattern of the equal-length sequence, and improves the enumeration method of the TS-FCPO algorithm to find the matching position in order to improve the speed of mining.The equal-length closed partial order inherits all the advantages of the closed partial order and can express the relation and deterministic trend of the closed sequence with the characteristics of equal length.Due to the fact that the equal-length partial order fully considers that only the corresponding position on the sequence can be matched between the sequences of equal length, the actual complexity of the algorithm can be reduced when it is applied to the application scene of the equal-length partial sequence.Time efficiency has been significantly improved.The results of this paper can be applied to the development of online book combination promotion plan, travel route planning and strategy.
【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】：TP311.13

【参考文献】