• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Advances in computational sciences and technology >Efficiently Maintaining and Discovering Sequential Patterns with Sequence Deletion using Discovered Sequences.
【24h】

Efficiently Maintaining and Discovering Sequential Patterns with Sequence Deletion using Discovered Sequences.

机译：使用发现的序列通过序列删除有效地维护和发现序列模式。

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Among the data mining techniques, sequential pattern mining is an important issue to discover the frequent subsequences from a sequential database. Maintaining sequential patterns should be paying more attention in dynamic databases, where the discovered sequences may become invalid when new sequences are added or deleted. In the past, Incremental Sequence Extraction (ISE) was proposed to maintain and discover sequential patterns with sequence insertion based on the discovered sequential patterns. Later FUSP tree structure and pre-large concepts were proposed to keep large and pre-large sequences (infrequent) to avoid the repeated scans over the original database for each update insertion and deletion. But the original database is required to be re-scanned for small sequences, which are necessary to be maintained in the updated database and required huge space for both. In this paper, we discuss the problem of maintaining and discovering sequential patterns when some not interested sequences are deleted. To overcome from the additional information, a candidate generate-test based approach MDSPU is presented. It makes full use of earlier mining result that is discovered sequences to speed up the process and cut down the cost of the process. The experimental results show that the proposed approach MDSPU is more than mining from scratch efficient in-terms of space and time.
机译：在数据挖掘技术中，顺序模式挖掘是从顺序数据库中发现频繁子序列的重要问题。维护顺序模式应该在动态数据库中得到更多的关注，在动态数据库中，添加或删除新序列后发现的序列可能变得无效。过去，提出了增量序列提取（ISE）来维护和发现序列模式，并基于发现的序列模式进行序列插入。提出了后来的FUSP树结构和预大概念，以保留大和预大序列（不频繁），以避免针对每次更新插入和删除重复扫描原始数据库。但是，需要对原始数据库进行重新扫描以查找较小的序列，这对于在更新的数据库中进行维护是必需的，并且这两者都需要巨大的空间。在本文中，我们讨论了删除一些不感兴趣的序列时维护和发现顺序模式的问题。为了克服附加信息，提出了一种基于候选生成测试的方法MDSPU。它充分利用了较早发现的发现结果序列，从而加快了流程并降低了流程成本。实验结果表明，所提出的方法MDSPU不仅仅是从零开始的高效时空方面进行挖掘。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号