本体标注和命名实体结合的传感器语义强化系统
发布时间：2018-12-26 07:22
【摘要】：物联网(Internet of Things,IoT)为了将传感器、控制器、使用者等一系列的物进行联系,这就需要一个标准的通信协议提供支持。通过物之间的联系,实现远程的管理控制以及智能化。在物联网中配置了大量的传感器,但这些传感器产生的数据多种多样且存在资源异构,将物联网中同一物体进行上传时,有可能会得到多种形式的表达。为解决机器不理解物的信息这一问题,在物联网中引入语义技术,形成语义物联网(Semantic Web of Things,SWoT)。针对以上不能正确地表达资源语义的问题,在语义物联网背景下,本文利用了本体和链接开放数据来表达语义信息,提出了一种本体标注和命名实体结合的传感器语义强化方法(Ontology Annotation and Named Entity combined Sensor Semantic Enhancement Method,OANESSEM)。该方法使用 SSN 本体标注传感器,为传感器数据添加语义,生成语义传感器数据,使机器理解传感器的语义;并采用命名实体相关方法,使用链接开放数据中的实体对传感器数据进行语义强化。在语义强化过程中,首先要从语义传感器数据中抽取命名实体,然后通过知识库DBpedia对其构建命名实体候选集,最后使用基于图的命名实体消歧方法,计算目标实体的最优解。语义强化的结果是从DBpedia中获取对命名实体描述最准确的目标实体URI,用于描述传感器,使用户更好的地理解传感器数据。最后依据该方法设计并开发了本体标注和命名实体结合的传感器语义强化系统。首先设计了系统的总体结构,并给出系统的第一级数据流图和各模块数据流图;然后给出系统运行过程图,并将实验结果与其它现有的方法进行比较。实验结果表明,本文所构建的语义强化系统可以较好地对传感器的语义进行描述,达到语义强化的目的,并且提高了查准率、查全率,可以较好地实现传感器语义的强化功能。
[Abstract]:Internet of things (Internet of Things,IoT) in order to connect sensors, controllers, users and so on, a standard communication protocol is needed. Through the connection between objects, the remote management and control and intelligence are realized. A large number of sensors are deployed in the Internet of things, but these sensors produce a variety of data and have heterogeneous resources. When uploading the same object in the Internet of things, it may be expressed in a variety of forms. In order to solve the problem that machines do not understand the information of objects, semantic technology is introduced into the Internet of things to form the semantic Internet of things (Semantic Web of Things,SWoT). In the context of semantic Internet of things, this paper uses ontology and link open data to express semantic information. In this paper, a new method of sensor semantic enhancement (Ontology Annotation and Named Entity combined Sensor Semantic Enhancement Method,OANESSEM) is proposed, which combines ontology tagging with named entities. In this method, SSN ontology is used to label the sensor, add semantics to the sensor data, generate the semantic sensor data, and make the machine understand the semantics of the sensor. Using the method of named entity correlation, the entities in the linked open data are used to enhance the semantic of sensor data. In the process of semantic enhancement, named entities are extracted from semantic sensor data first, then named entity candidate sets are constructed by knowledge base DBpedia. Finally, the optimal solution of target entities is calculated by using graph-based named entity disambiguation method. The result of semantic enhancement is that the target entity URI, which is the most accurate description of named entity, is obtained from DBpedia to describe the sensor, so that the user can understand the sensor data better. Finally, a sensor semantic enhancement system based on ontology annotation and named entity is designed and developed. First, the overall structure of the system is designed, and the first stage data flow diagram and each module data flow diagram of the system are given, then the running process diagram of the system is given, and the experimental results are compared with other existing methods. The experimental results show that the semantic enhancement system constructed in this paper can better describe the semantic of the sensor, achieve the purpose of semantic enhancement, and improve the precision and recall rate, which can better realize the semantic enhancement function of the sensor.
【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】：TP212;TP391.44;TN929.5
[Abstract]:Internet of things (Internet of Things,IoT) in order to connect sensors, controllers, users and so on, a standard communication protocol is needed. Through the connection between objects, the remote management and control and intelligence are realized. A large number of sensors are deployed in the Internet of things, but these sensors produce a variety of data and have heterogeneous resources. When uploading the same object in the Internet of things, it may be expressed in a variety of forms. In order to solve the problem that machines do not understand the information of objects, semantic technology is introduced into the Internet of things to form the semantic Internet of things (Semantic Web of Things,SWoT). In the context of semantic Internet of things, this paper uses ontology and link open data to express semantic information. In this paper, a new method of sensor semantic enhancement (Ontology Annotation and Named Entity combined Sensor Semantic Enhancement Method,OANESSEM) is proposed, which combines ontology tagging with named entities. In this method, SSN ontology is used to label the sensor, add semantics to the sensor data, generate the semantic sensor data, and make the machine understand the semantics of the sensor. Using the method of named entity correlation, the entities in the linked open data are used to enhance the semantic of sensor data. In the process of semantic enhancement, named entities are extracted from semantic sensor data first, then named entity candidate sets are constructed by knowledge base DBpedia. Finally, the optimal solution of target entities is calculated by using graph-based named entity disambiguation method. The result of semantic enhancement is that the target entity URI, which is the most accurate description of named entity, is obtained from DBpedia to describe the sensor, so that the user can understand the sensor data better. Finally, a sensor semantic enhancement system based on ontology annotation and named entity is designed and developed. First, the overall structure of the system is designed, and the first stage data flow diagram and each module data flow diagram of the system are given, then the running process diagram of the system is given, and the experimental results are compared with other existing methods. The experimental results show that the semantic enhancement system constructed in this paper can better describe the semantic of the sensor, achieve the purpose of semantic enhancement, and improve the precision and recall rate, which can better realize the semantic enhancement function of the sensor.
【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】：TP212;TP391.44;TN929.5
【相似文献】
相关期刊论文 前10条
1 赵君U，
本文编号：2391727
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2391727.html
