Research And Implementation Of Hybrid Similarity Algorithm Based On OWL-S

Web services are service-oriented model of distributed computing. Web services has a good distribution, low coupling, and pluggable, and programmers only need to know the service interface, without having to know the details of implementation, which makes web applications development, maintenance and updating of the efficiency greatly enhanced. Web service discovery is an important part of the web service system. Web service discovery aims to meet the requirements of the service from a large number of web services. Because of the growing number of web services, questions are followed by. The traditional web services protocol is lack of good definition of semantic information, and web services can only support the simple matching of words, which cause a low recall rate and precision. Because of interface is based on the grammar, the user or program can not get more information about service functions, which bring about selecting accurately web services difficulty. The emergence of semantic web provides solutions for the above-mentioned problems, describing the semantics of service with the knowledge of the semantic web markup. You can make web services understood and reasoned by computer, which can evaluate correlation between service requests and advertising services from the semantic level. Semantic web services can be flexibly and reliably provide web services in accordance with the needs of users in dynamic, heterogeneous web environment. Semantic web service discovery research includes a description of the semantic web service, discovery architecture and the algorithm of semantic web service discovery. These three aspects interrelate and affect the recall rate and precision of the semantic web service discovery. However, the technology of semantic web service discovery is not maturity, and the two main questions are the following: first is the description of the semantic web service; second is the matching algorithm based on semantic.In this paper, in order to improve the semantic web service discovery, we will carry out in-depth study from the main problems by the study and comparison of the current technology at home and abroad on the basis of the current web service discovery. The research is focused by the following points:Firstly, through studying ontology technology for web services matching, realize the ontology-based web service discovery mechanism. Ontology is used to describe the relation between resources. Through the definition of the semantics used by ontology, machines can interoperate and understand the semantics of data. The abstract of Ontology model are described as following: D = , C represents the concept collections. R ? C?C represents collections of the relationship between concepts, such as the equivalence relations, inheritance relations and inclusion relations.Secondly, analysis of the web service description language OWL-S. The profile will be focused because service discovery mostly use OWL-S's Profile. Profile typically includes the provision of services, the description of the service function and the description of the service characteristic. The provisions of services provide a description of the information available for manual reading. The description of the service function is about that which kinds of function the user provide and which kinds of conditions description about successful outcome is to be needed. The description of the service characteristic is about additional features of the service. In this paper, the model will use the the description of provisions of services and the service function.Thirdly, at the basis of current semantic matching algorithms and some problems existence, a hybrid similarity method for service discovery based on the OWL-S is raised. This algorithm is the main innovation of this paper. This algorithm is a hybrid of the distance-based similarity algorithm and the IR-based cosine similarity algorithm. The IR-based cosine similarity algorithm inherits the text similarity algorithm, taking into account the content of information, but the algorithm does not make full use of semantic information for reasoning. Similarity algorithm based on the distance greatly improves the classical algorithm and introduces the binary relationship between the concepts. This algorithm uses semantic information, but not takes into account the content of information. Hybrid measurement method uses mutli-standards in the process. Comparing with a number of measurement methods executing separately, it can provide better candidates and better performance results, so this paper will use the hybrid similarity algorithm mixing the two methods and learning from their strengths. Taking into account the semantic information and the content information, you can fully exploit the service of information resources to achieve a better match between semantic web services and improve the rate of recall and precision.Finally, combination of the above study, we designed a model of semantic web service discovery system. This system includes service query, OWL-S service description database, semantic services discovery engine and ontology database. The function of service query is an user interface for interactive; OWL-S service description database stores and reads web services which are described by OWL-S description; Ontology database stores ontology needed in matching process. The ontology provides the concept definition and the relationship between concepts, which provide a unified abstraction ontology glossary for services ontology. Semantic web service matchmaking engine is the core module of the system, including reasoning and matching device. Semantic reasoning device is based on OWL and description logic of semantic equivalence. Through the using of description logic with containing relationship between concepts, determine and reason relationship between concepts of ontology. The functions of matching device make input and output of the semantic information which the reasoning engine provides and parameters of information as a basis for matching. According to the algorithm, matchs the two sides of the IO, screens services by the minimum similarity which service requestor set, sorts results according to the magnitude of similarity and returns the optimal solution to the user.This structure implements the discovery of services, provides the semantic reasoning ability and achieves good discovery results. Through an experimental example, the model of the design is made more comprehensive performance analysis, and tests its characteristics and the ability of matching algorithm. The purposes of improving semantic web discovery are achieved.In this paper, through adopting lots of the current technology at home and abroad, results of my work successfully improve the semantic web service discovery and have some reference value for current research in this field.