跨系统的流动性多源异构数据整合算法仿真  

Simulation of Mobile Multi-Source Heterogeneous Data Integration Algorithms Across Systems

作　　者：陈亚颐 尹权 再开日亚·安尼娃尔 张乐[2] CHEN Ya-yi;YIN Quan;Zai kairiya·Anniwaer;ZHANG Le(Xi'an Shiyou University,Xi'an Shaanxi 710065,China;Xinjiang Oilfield Company Zhundong Oil Production Plant,Fukang Xinjiang 831500,China)

机构地区：[1]西安石油大学,陕西西安710065 [2]新疆油田公司准东采油厂,新疆阜康831500

出　　处：《计算机仿真》2025年第1期410-414,共5页Computer Simulation

基　　金：2023年准东采油厂公共光缆维护项目(XYJL-2023-FW-062)。

摘　　要：多源异构数据的存在,使数据在不同系统之间的流动和共享变得复杂而困难,导致数据资源无法被充分利用,形成了数据孤岛。为了提高数据质量和利用效率,提出跨系统流动性的多源异构数据整合算法研究。利用时序关联和密度聚类算法对收集到的跨系统流动性多源异构数据实施数据清洗,提高数据质量;采用堆叠自编码器深度神经网络(Stacked Auto-Encoder, SAE)从跨系统数据源中抽取出描述数据跨系统流动性的关键特征点;建立基于循环神经网络的数据整合模型,将这些关键特征点作为输入,并通过该模型不断优化,实现跨系统的多源异构数据高效整合。实验结果表明,所提方法得到的数据具有较高的质量,且相似度控制在0.895~0.960之间,整合效果最为可靠。The existence of multi-source heterogeneous data makes the flow and sharing of data between different systems complicated and difficult,which leads to the inability to make full use of data resources and forms data islands.In order to improve data quality and utilization efficiency,a multi-source heterogeneous data integration algorithm with cross-system liquidity is proposed.Time series correlation and the density clustering algorithm are used to clean the collected cross-system liquidity multi-source heterogeneous data to improve the data quality.The Stacked Auto-Encoder,SAE is used to extract the key feature points that describe the cross-system fluidity of data from cross-system data sources.A data integration model based on circular neural networks is established.These key feature points are taken as input,and the model is continuously optimized to realize efficient integration of multi-source heterogeneous data across systems.The experimental results show that the data obtained by the proposed method has high quality,and the similarity is controlled between 0.895 and 0.960,so the integration effect is the most reliable.

关 键 词：跨系统流动性 多源异构数据整合 数据清洗 特征抽取 循环神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]