Research On A Virtual Web Service Method Combining MDS And Gaussian LDA

At present,the number of Web services in the Internet is increasing day by day,and it is not easy for users to obtain the required Web services accurately and efficiently.Service discovery technology aims to provide a series of optional services according to user needs to improve the efficiency of Web service use,and virtual Web service technology aims to shield the differences generated when calling services and improve the quality of Web services.It is particularly important to combine the two reasonably.In general,the pain points of service discovery technology are mainly low efficiency and low accuracy.This paper proposes a high-accuracy Web service clustering method as a prerequisite for service discovery,and then solves these two problems.The main pain point of virtual Web service technology is the difficulty of unified management.On the premise of obtaining service discovery results,this paper uses microservice technology to solve this problem.This paper proposes improved algorithms for service clustering,service discovery,and service virtualization in the field of service computing.The specific work and innovations are as follows:(1)At present,the popular Web services used in the industry are API-based Web services described in natural language,and the existing virtual Web service research pays less attention to this.Aiming at such Web services,this paper designs a virtual Web service system that integrates service clustering,service discovery,and load balancing.(2)Aiming at the problem that it is difficult to extract features from Web service description documents,resulting in poor clustering accuracy,this paper uses natural language processing technology to extract service functions,and then combines the ontology theory to compare the differences between service functions,and then uses multi-dimensional scaling analysis to calculate the service function vector Then use Gaussian latent Dirichlet distribution for text modeling to extract semantic features of service description documents to achieve service clustering.More accurate service clustering results are an important basis for subsequent service discovery.(3)In the process of service discovery,this paper proposes a method to incorporate new elements into the existing fitting composition when dealing with multi-dimensional scaling analysis models.Combined with the big top heap,the pruning search is performed based on the service clustering results,which improves the efficiency and accuracy of service discovery.(4)This paper uses microservice technology to register and manage Web services,and realizes automatic service discovery and virtualized calls through the description information input by users.And a load balancing algorithm based on response time differential sequence is proposed,which improves the performance of virtual Web services in high concurrency scenarios.All the algorithms in this paper give a detailed flow and derivation process,and conduct experiments with real Web services in programmable Web.The experimental results show that the method proposed in this paper can significantly improve the corresponding indicators in service clustering,service discovery,and load balancing.