Web服务网络分析和社区发现研究
发布时间：2018-11-05 11:47
【摘要】：当前服务计算(SOC)领域的研究重点关注于Web服务的发布、注册、组织、管理，以及通过服务发现、组合等SOC任务满足终端用户的需求。然而随着互联网的发展和各种计算模式的兴起给SOC带来了新的机遇和挑战，一方面，互联网上的服务正在向着大规模、大数据量、多样性、复杂性的方向发展，逐渐形成了以Web服务为中心的复杂系统；另一方面，由于服务之间不能彼此感知，以及缺乏有力的语义支持，使得互联网上的服务存在着信息孤岛问题。针对上述挑战，本文从服务互联分析和社区发现的角度对大规模服务复杂系统进行相关研究。 本文是综合了语义Web技术、复杂网络技术、主题建模技术对服务计算进行的交叉研究。其中语义Web技术是传统的，也是当前服务计算研究的重点，它的主要目的是使得服务具有语义，从而使得计算机能够理解服务、自动化的处理服务；而传统的语义Web技术主要通过借鉴外部知识库(如本体)实现对服务的语义标注，这就可能导致标注不精确以及本体对服务领域多样性的适用问题，本文通过借助主题建模技术学习服务本身蕴含的语义信息来进一步增强服务的语义；复杂网络技术主要针对服务的复杂系统问题，从整体上分析面向服务的复杂系统的结构、规律、社区等。本文主要贡献如下： 1.面向Web服务的复杂网络分析和优化 服务形成的网络可以有效表达服务之间的互操作关系，而这些互操作是潜在的服务组合模式。基于网络化的服务结构，自动服务组合的过程可以通过图搜索来完成，从而提高组合的效率。基于服务之间的互操作关系，本文自底向上的构建了面向语义Web服务的两种网络模型，并针对网络中的孤立节点和不合理链路问题给出了解决方案：通过借鉴网络分析，预测一些对当前网络具有重要价值的服务，从而为增强服务组合网络的连通性提供支持；另一方面，针对不合理链路，给出一种潜在的基于重用用户历史组合案例的组合网络优化机制； 2.面向网络化Web服务的社区结构挖掘 通过研究以Web服务为中心的社区结构，能够揭露网络化服务中深层次的动态结构，从而为服务的组织、管理，以及服务发现和组合中快速、准确的定位满足需求的服务提供有效的手段。根据第一部分的网络模型，，探讨了两种面向网络化服务的社区结构：根据服务功能参数在本体中的语义关系，面向功能语义的社区结构揭示了具有相似功能的服务的聚集和竞争现象；基于服务与服务之间互操作关系，构建并分析了面向协作的社区结构。最后分析了这两种社区结构在精化用户需求和设计高效组合应用上具有较高的价值； 3. Web API/Mashup网络分析及挖掘 针对WebAPI及Mashup形成的软件生态系统进行自顶向下的网络建模和复杂网络特性分析。通过构建多种网络模型对WebAPI及Mashup的功能语义特性以及集成特性进行了全面挖掘，分析了所构建网络的小世界特性、幂律特性以及小集团结构，并从实际应用的角度探讨了这些发现能够帮助用户、服务提供者有效地利用WebAPI及Mashup生态系统，如探索式、交互式的用户需求精化； 4.面向服务的主题建模及分析 针对传统服务语义标注不精确性，以及本体对于领域多样性不适用的问题，探讨了主题建模技术对服务计算的价值。针对经典的WSDL Web服务进行了LDA挖掘，给出了LDA在服务聚类和社区划分中的应用；其次探讨了本体对监督式LDA的借鉴意义，并提出基于本体的Labeled LDA扩展模型OEL-LDA，通过实验对比验证了所提模型对Labeled LDA中主题质量的有效提高，并通过实验给出了融合本体对精化Labeled LDA中冗余主题的必要性；
[Abstract]:Research in the field of current service computing (SOC) focuses on publishing, registering, organizing, managing Web services, and meeting end-user needs through SOC tasks such as service discovery, combination, and so on. However, as the development of the Internet and the rise of various computing models bring new opportunities and challenges to the SOC, on the one hand, the services on the Internet are developing towards large scale, large data volume, variety and complexity, and gradually form a complex system centered on Web services. On the other hand, because services cannot be perceived to each other, and lack of strong semantic support, there is an island problem of information on the Internet. Aiming at the above challenge, this paper studies the complex system of large-scale service from the view of service interconnection analysis and community discovery. This paper deals with the intersection of semantic Web technology, complex network technology and topic modeling technology on service calculation. Research. Semantic Web technology is traditional and is the focus of the current service computing research. Its main purpose is to make the service have semantic, so that the computer can understand the service and automation. The traditional semantic Web technology can be used to mark the semantic annotation of the service by referring to the external knowledge base (such as ontology), which can lead to inaccurate marking and the application of ontology to the service field. In this paper, the semantics of service is further enhanced by learning the semantic information contained in the service itself with the help of the topic modeling technology, the complex network technology mainly focuses on the complex system problem of the service, the structure, the law and the society of the complex system facing the service are analyzed from the whole. Area, etc. The main contribution of this article The following: 1. Complex Network for Web Services The networks formed by the network analysis and optimization service can effectively express the interoperation relationship between services, and these interoperation are latent in a networked-based service architecture, the process of the automated service combination may be accomplished through the graph search, from In order to improve the efficiency of the combination, this paper constructs two kinds of network models facing semantic Web service from the bottom up, and gives the solution to isolate nodes and unreasonable links in the network. By using the network analysis, we can predict some services which have important value to the current network, so as to provide support for the connectivity of the enhanced service combination network; on the other hand, aiming at the unreasonable link, a potential reuse user history combination case is given. Combinatorial network optimization mechanism; 2. network oriented Web service community structure mining can expose the deep dynamic structure in the network service by studying the community structure centered on Web service, thus serving the organization and the pipe of the service. Quick and accurate positioning in discovery and combination of services According to the network model of the first part, two community structures for networked services are discussed: the semantic relation of the service function parameters in the ontology, and the community structure facing the functional semantics reveals the phase of the community structure. The aggregation and competition of functional services; interoperation between services and services. The collaborative community structure was built and analyzed. Finally, the requirements and design of the two community structures in refined users were analyzed. High-efficiency combined applications with higher value; 3. 1.3. 1. b API/ Mashup Network Analysis and Mining Software Ecosystem for WebAPI and Mashup In this paper, a top-down network modeling and complex network characteristic analysis are carried out. By constructing a variety of network models, the function semantic characteristics and integration characteristics of WebAPI and Mashup are comprehensively mined, and the small world characteristics, power law characteristics and small group structure of the constructed network are analyzed. The perspective of practical applications explores these findings to help users and service providers effectively leverage the WebAPI and Mashup Ecosystem, such as exploration, interactive User Requirements refinement; 4. Service-oriented topic modeling and analysis of traditional services The semantic annotation is not accurate, and the ontology is much in the field. This paper discusses the value of the subject modeling technology to the service calculation, discusses the application of the ontology to the service and community partition, and then discusses the significance of the ontology to the supervised classification, and puts forward some suggestions. Based on the body-based Labeled Dirichlet extended model OEL-044, the effectiveness of the proposed model on the quality of the subject in Labeled has been verified by experimental comparison, and the experimental results are given.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP393.09
本文编号：2311994
[Abstract]:Research in the field of current service computing (SOC) focuses on publishing, registering, organizing, managing Web services, and meeting end-user needs through SOC tasks such as service discovery, combination, and so on. However, as the development of the Internet and the rise of various computing models bring new opportunities and challenges to the SOC, on the one hand, the services on the Internet are developing towards large scale, large data volume, variety and complexity, and gradually form a complex system centered on Web services. On the other hand, because services cannot be perceived to each other, and lack of strong semantic support, there is an island problem of information on the Internet. Aiming at the above challenge, this paper studies the complex system of large-scale service from the view of service interconnection analysis and community discovery. This paper deals with the intersection of semantic Web technology, complex network technology and topic modeling technology on service calculation. Research. Semantic Web technology is traditional and is the focus of the current service computing research. Its main purpose is to make the service have semantic, so that the computer can understand the service and automation. The traditional semantic Web technology can be used to mark the semantic annotation of the service by referring to the external knowledge base (such as ontology), which can lead to inaccurate marking and the application of ontology to the service field. In this paper, the semantics of service is further enhanced by learning the semantic information contained in the service itself with the help of the topic modeling technology, the complex network technology mainly focuses on the complex system problem of the service, the structure, the law and the society of the complex system facing the service are analyzed from the whole. Area, etc. The main contribution of this article The following: 1. Complex Network for Web Services The networks formed by the network analysis and optimization service can effectively express the interoperation relationship between services, and these interoperation are latent in a networked-based service architecture, the process of the automated service combination may be accomplished through the graph search, from In order to improve the efficiency of the combination, this paper constructs two kinds of network models facing semantic Web service from the bottom up, and gives the solution to isolate nodes and unreasonable links in the network. By using the network analysis, we can predict some services which have important value to the current network, so as to provide support for the connectivity of the enhanced service combination network; on the other hand, aiming at the unreasonable link, a potential reuse user history combination case is given. Combinatorial network optimization mechanism; 2. network oriented Web service community structure mining can expose the deep dynamic structure in the network service by studying the community structure centered on Web service, thus serving the organization and the pipe of the service. Quick and accurate positioning in discovery and combination of services According to the network model of the first part, two community structures for networked services are discussed: the semantic relation of the service function parameters in the ontology, and the community structure facing the functional semantics reveals the phase of the community structure. The aggregation and competition of functional services; interoperation between services and services. The collaborative community structure was built and analyzed. Finally, the requirements and design of the two community structures in refined users were analyzed. High-efficiency combined applications with higher value; 3. 1.3. 1. b API/ Mashup Network Analysis and Mining Software Ecosystem for WebAPI and Mashup In this paper, a top-down network modeling and complex network characteristic analysis are carried out. By constructing a variety of network models, the function semantic characteristics and integration characteristics of WebAPI and Mashup are comprehensively mined, and the small world characteristics, power law characteristics and small group structure of the constructed network are analyzed. The perspective of practical applications explores these findings to help users and service providers effectively leverage the WebAPI and Mashup Ecosystem, such as exploration, interactive User Requirements refinement; 4. Service-oriented topic modeling and analysis of traditional services The semantic annotation is not accurate, and the ontology is much in the field. This paper discusses the value of the subject modeling technology to the service calculation, discusses the application of the ontology to the service and community partition, and then discusses the significance of the ontology to the supervised classification, and puts forward some suggestions. Based on the body-based Labeled Dirichlet extended model OEL-044, the effectiveness of the proposed model on the quality of the subject in Labeled has been verified by experimental comparison, and the experimental results are given.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP393.09
【参考文献】
相关期刊论文 前2条
1 邓水光;吴朝晖;;Web服务组合方法综述[J];中国科技论文在线;2008年02期
2 杨春芳;;圆顶阀在固体物料传输中的应用[J];中国仪器仪表;2010年10期
本文编号：2311994
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2311994.html
