基于改进RNN多源融合算法的网络异构信息集成管理系统  

Network heterogeneous information integrated management system based on improved RNN multi-source fusion algorithm

作　　者：李麟 王伟 LI Lin;WANG Wei(Information Department of the Navy Qingdao Special Service Rehabilitation Center,Qingdao 266071,Shandong,China)

机构地区：[1]海军青岛特勤疗养中心信息科,山东青岛266071

出　　处：《西安工程大学学报》2023年第6期145-152,共8页Journal of Xi’an Polytechnic University

基　　金：海军官兵区域康复医疗保障体系构建及技术研究(20BJZ08);海军青岛特勤疗养中心科研任务书(TLKY08)。

摘　　要：医疗信息网络的互联造成其网络架构极为复杂,而当前对于此类网络产生的多源异构数据的管理水平难以满足信息化的要求,故给出了一种改进的信息集成管理系统。文中采用野马优化(wild horse optimizer, WHO)算法改进的循环神经网络(recurrent neural network, RNN)设计多源异构数据融合模型,并在此基础上以生产执行系统(manufacturing execution system, MES)为核心建立网络异构信息集成管理系统。通过同步集成处理多源融合模型输出的数据,将所有数据传递至系统数据库中进行储存,并通过所设计的功能模块完成了多源异构数据的集成管理。实验结果表明:以WHO-RNN多源融合算法为核心的网络异构信息集成管理系统的数据融合完整性维持在0.700~0.800的较高水平,并且随着数据量增加,末态更新水平(ultimate update level, UUL)指标逐渐趋近于80%,系统对于多源异构数据的运算效率和识别精度较好,具有较强的工程实际应用性能。The interconnection of medical information networks results in extremely complex net-work architectures,and the current management level of multi-source heterogeneous data genera-ted by such networks is difficult to meet the requirements of informatization.Therefore,an im-proved information integration management system is proposed.The article adopted the wild horse optimizer(WHO)algorithm to improve the recurrent neural network(RNN)and designed a multi-source heterogeneous data fusion model.Based on this,a network heterogeneous infor-mation integration management system was established with the manufacturing execution system(MES)as the core.By synchronously integrating and processing the data output from the multi-source fusion model,all data was transmitted to the system database for storage,and the inte-grated management of multi-source heterogeneous data was completed through the designed func-tional modules.The experimental results show that the data fusion integrity of the network het-erogeneous information integration management system with WHO-RNN multi-source fusion al-gorithm as the core is maintained at a high level of 0.700~0.800,and as the data volume increa-ses,the ultimate update level(UUL)index gradually approaches 80%.The system has great computational efficiency and recognition accuracy for multi-source heterogeneous data,and strong engineering practical application performance.

关 键 词：循环神经网络(RNN) 野马优化(WHO)算法 生产执行系统(MES) 多源异构数据 数据融合 信息集成管理系统 医疗信息网络 

分 类 号：TP315[自动化与计算机技术—计算机软件与理论]