语义物联网中基于上下文感知的服务选择方法

发布时间：2019-05-23 15:30

【摘要】：物联网(Internet of Things)是一个由各种智能传感设备组成的庞大网络。由于物联网信息的不完备性和不确定性阻碍了物联网中信息协调和交互,将语义技术引入物联网中,从而产生了语义物联网(Semantic Web of Things,SWoT)。语义物联网服务特征表现为海量性、开放性、动态性、不完全性和异构性,导致难以根据用户需求选择服务。为了解决这一问题,本文提出了一种语义物联网中基于上下文感知的服务选择方法。充分考虑用户选择时的语义、显性偏好、隐性偏好和上下文信息。首先,对于服务库中信息的模糊性和用户表达的不明确性,进行语义标注和分类使得信息更容易协同和交互,选择出满足用户需要的服务。其次,获得用户的显性和隐性偏好,选择出满足用户显性和隐性偏好的候选服务集。其中,显性偏好是由用户提出的需求直接获得,隐性偏好则由传感器所获得的用户上下文信息与模糊规则库经过推理机推理得到。再次,将候选服务集进行量化(离散化)和归一化处理,其结果作为萤火虫优化神经网络的输入,选择最佳萤火虫来确定神经网络的权重与阈值以提高训练效率,最终为用户返回有序的服务列表。为验证方法的可行性,本文设计了语义物联网中基于上下文感知的服务选择原型系统(SWoT-oriented context aware-based service selection prototype system,SWoT-CBSS)。该系统从功能上分为服务语义标注、服务分类、服务筛选和萤火虫优化神经网络四个模块,分别实现了语义标注、分类与选择功能。通过用户选择酒店服务对系统进行验证。结果表明,本文设计的SWoT-CBSS在查准率和查全率都有一定的保证,虽然响应时间略长,但总体上能够得到满足用户需求的服务。
[Abstract]:Internet of things (Internet of Things) is a huge network composed of various intelligent sensing devices. Because the incompleteness and uncertainty of Internet of things information hinder the coordination and interaction of information in the Internet of things, semantic technology is introduced into the Internet of things, resulting in semantic Internet of things (Semantic Web of Things,SWoT). The semantic Internet of things service is characterized by sea quantity, openness, dynamics, incompleteness and heterogeneity, which makes it difficult to choose services according to the needs of users. In order to solve this problem, this paper proposes a context-aware service selection method in semantic Internet of things. Fully consider the semantics, explicit preferences, implicit preferences and contextual information of user selection. First of all, for the fuzziness of information and the uncertainty of user expression in the service library, semantic tagging and classification make the information easier to cooperate and interact, and select the services to meet the needs of users. Secondly, the explicit and implicit preferences of users are obtained, and the candidate service sets that meet the explicit and implicit preferences of users are selected. Among them, the explicit preference is obtained directly by the requirements proposed by the user, and the implicit preference is obtained by the inference engine reasoning of the user context information and fuzzy rule base obtained by the sensor. Thirdly, the candidate service set is quantified (discretized) and normalized, and the results are used as the input of the firefly optimization neural network, and the optimal firefly is selected to determine the weight and threshold of the neural network in order to improve the training efficiency. Finally, an orderly list of services is returned for the user. In order to verify the feasibility of the method, a context-aware service selection prototype system (SWoT-oriented context aware-based service selection prototype system,SWoT-CBSS) in semantic Internet of things is designed in this paper. The system is divided into four modules: service semantic tagging, service classification, service screening and firefly optimization neural network. The functions of semantic tagging, classification and selection are realized respectively. The system is verified by the user's choice of hotel service. The results show that the SWoT-CBSS designed in this paper has a certain guarantee in precision and recall. Although the response time is a little longer, it can get the service to meet the needs of users as a whole.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.44;TN929.5

【相似文献】