摘要
针对当前数据更新方法中存在实时性和安全性差的问题,提出基于面向对象的分布式数据库关键信息增量式更新方法。以数据库中关键信息样本增加后条件熵的变化情况和对约简集合影响规律为依据,得到新增信息数据样本不一致邻域导致条件熵发生变化,且约简集合也产生相应变化的结论。从信息观角度出发,对新增加的信息数据样本和其不一致邻域实行约简操作,实现更新变动之前邻域数据约简。基于约简结果,将信息数据增量分为增加与删除两部分,通过更新前后相同数量数据进行叠置与比较,识别出增量数据。通过逐级条件筛选方式将识别出的增量数据进行配准,实现数据融合。根据增量识别与增量融合,通过面向对象理念同步处理分布式数据中增量数据,完成数据库关键信息增量式更新。实验结果表明,所提方法数据更新效率高、安全性强。与当前相关方法比较,该方法鲁棒性更强,更具可靠性。
In this article, an incremental update method for key information in distributed databases based on oriented object was proposed. Based on the change of conditional entropy and the influence law of reduction set after the increase of key information samples in database, the inconsistent neighborhood of new information data sample led to the change of conditional entropy, and the reduction set also caused corresponding changes. From the information perspective, the additional information data sample and its inconsistent neighborhood were reduced, so that the neighborhood data reduction before the update was achieved. Based on the reduction result, the information data increment was divided into two parts: the addition and the deletion. Then, the incremental data was identified by superimposing and comparing the same amount of data before and after the update. The identified incremental data was registered by step-by-step condition filtering, so that the data fusion was achieved. According to the incremental identification and incremental fusion, the incremental data in distributed data was processed synchronously through the object-oriented idea. Finally, the incremental update of key information was completed. Simulation results show that the proposed method has high data update efficiency and strong security. Compared with the current methods, this method is more robust and more reliable.
作者
王乐
宋利敏
WANG Le;SONG Li-min(Sias International University,Zhengzhou University,Xinzheng Henan 451100,China)
出处
《计算机仿真》
北大核心
2020年第6期416-419,共4页
Computer Simulation
基金
2018大学生实验创新项目(2018SYCX037)。
关键词
分布式数据库
关键信息
增量式
更新
Distributed database
Key information
Incremental
Update
