基于完井业务流程的加权序列模式挖掘算法研究  

Weighted sequential pattern mining algorithm research based on well completion business process

作　　者：尚福华[1] 孙姝凝 陈效果 杜睿山[1] 

机构地区：[1]东北石油大学计算机与信息技术学院,黑龙江大庆163318 [2]东方地球物理勘探有限公司,河北保定072750

出　　处：《计算机应用研究》2014年第9期2719-2723,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61170132);国家重大专项资助项目(2011ZX05020-007);黑龙江省教育厅科学技术研究资助项目(12521055)

摘　　要：针对完井移动平台Web访问模式中用户对整体上符合完井业务流程习惯的序列模式更加感兴趣的特点,提出一种基于完井业务流程的加权序列模式挖掘算法。通过对完井业务流程模型和完井Web访问日志作分析,确定完井业务依赖强度计算模型并作为序列的权重,同时在加权挖掘的过程中采用k-最小加权支持数技术,对PrefixSpan算法进行改进。该算法在加权挖掘的过程中可以舍弃一大批不满足业务流程的访问序列,有效地避免候选组合爆炸的发生。实验表明该算法能快速地挖掘出有意义的完井访问序列模式。Focusing on the users are more interested in the sequential patterns that accord with the well completion business process habit overall in the Web access mode of well completion mobile platform, this paper proposed a weighted sequential pattern mining algorithm based on the well completion business process. Through analyzing the business process models and Web access log of well completion, it confirmed the business dependency strength calculation model of well completion as the sequence weight, at the same time, using the technology of k-minimum weighted support in the weighted mining to improve the PrefixSpanalgorithm. The algorithm discards a lot of access sequence that cannot meet the needs of the business process in the weighted mining, avoids the happening of the candidate combination explosion effectively. Experiments prove that the algo- rithm can rapidly excavate meaningfulwell completion sequential patterns.

关 键 词：数据挖掘 完井业务流程 加权序列模式 最小加权支持数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]