Research On Web API Recommendation Method Based On Mashup Service Function Semantic Clustering

Mashup technology is one of the key technologies of Web 2.0.It can integrate a variety of Web APIs with different functions to build a brand new Web application called Mashup service.With the support of Mashup technology,many combination-level applications can complete functional integration and data integration in a short period of time,which greatly improves the efficiency of application development,and is therefore favored by many enterprises and technical institutions.However,a large number of Mashup services and Web APIs are registered in mainstream service warehouses.How to quickly and accurately locate Web APIs that meet the needs of Mashup in such a large set of services has become an urgent problem.For this reason,this paper focuses on three key issues of feature representation,service clustering,and Web API recommendation of Mashup service functions.The main research contents include:(1)This paper proposes a feature vector representation method based on Mashup service function semantics.This method first normalizes the Mashup service description text,and then extends the Mashup service label according to the invocation relationship between the Mashup service and the Web API to ensure that the label is relatively reasonable and complete.Based on this,a functional semantic association calculation method(FSAC)is proposed,which uses the service tag information and functional noun information in the service description to perform semantic association calculation,and integrates the calculation result into the construction process of Mashup semantic feature vector,so that it can better reflect the true functional characteristics of the service.(2)This paper proposes a Mashup service clustering method based on density peak detection.This method first redefines the local density calculation method in the traditional density peak clustering algorithm,which makes the density distribution and related density information calculation more reasonable.Based on this,a clusteringcenter detection method based on density information(CCD-DI)is proposed.This method is used to select the optimal K cluster centers as the input of the K-means algorithm to implement the Mashup service clustering,thereby ensuring the accuracy and stability of the clustering.(3)This paper proposes a Web API recommendation algorithm based on Mashup service neighborhood(WR-MSN).The algorithm first expresses the Mashup service requirements in the form of requirement semantic feature vector.Then,based on the clustering result of the Mashup service,the algorithm locks the Mashup service neighborhood that best matches the requirement semantic feature vector.Based on the invocation relationship between Mashup service and Web API in the Mashup service neighborhood,the algorithm divides the functional categories of Web API data and ranks the Web API data of each category.In this way,the Web API combination that best matches Mashup service requirements is selected for recommendation.