Research On Web Service Discovery Method Based On Topic Model

As a kind of autonomous and open application entity,Web service is a new distributed computing model.It has the characteristics of loosely coupled,platform independent and interoperable.It is especially suitable for publishing and using in Internet environment.Web service discovery technology is an important part of Web services architecture.Currently,there are mainly keyword-based and semantic-based methods for Web service discovery.Keyword-based service discovery can not understand user semantic information,which leads to low recall.The traditional method based on semantic service discovery is too limited because of the limitation of promotion.The topic-based service discovery method is also a kind of semantic discovery method.Compared with the traditional semantic discovery method,it has the characteristics of less restrictive condition and strong generalization.Based on the existing work,this paper presents Web service discovery method based on topic model and clustering,which is based on Web services discovery and topic model.Firstly,according to the characteristics of WSDL description document,combined with the domestic and foreign text data processing methods,the feature extraction of the document,removal of stop words and tags,conjunction word segmentation,capital letter conversion,stemming,etc.,get the data processing document vocabulary collection.Then,the theme is modeled based on the BTM data vocabulary collection,we use Gibbs sampling to train the subject,and the theme number is determined by calculating the theme structure similarity to get the topic information of the document.Next,we study the method of text similarity measurement,and give a way to calculate the similarity between service-topic vector and service-word weight vector linearly.In order to improve the computing efficiency,we use the k-means algorithm based on the maximum distance method to cluster the Web services,and get the clustering of Web services.Finally,when querying the Web service,find the most relevant Web service cluster for the query,and use the Web service with high relevance in the cluster as the result of the discovery.This paper implements the specific modules in the method.Finally,the experimental results show that the method proposed in this paper has a high precision of Web service discovery.And the research of this paper has the reference significance to the related processing of Web service discovery.