The Research And Implementation Of Mashup Platform Based On RESTful WEB Services

With the rapid development of the Internet,Web service,a simple and easy way to use Internet applications,has been widely used in all aspects of the Internet.Now RESTful Web Services which are easy to use are becoming the main form to open services.Meanwhile,there are more and more advanced applications based on Web services,such as Mashup,a new technology of service composition.However,researches and works based on RESTful services are currently few.The main purposes of this thesis are:first is to study service matching and service recommendation,both are popular technologies of Web services,second is to develop a Mashup platform based on REST services and implement these technologies as important functions on it.The thesis studies service matching firstly.After introducing a service description for REST services,algorithms based on features of several important parts are proposed.Compared to some matching method proposed by now,this method covers more service features and it can achieve higher precision and recall.Then,the thesis studies service recommendation and the idea of collaborative filtering is discussed mainly.Firstly,considering the requirements of service recommendation from Mashup platform,a new method of user similarity is proposed.It uses a new mesure method for user similarity and can cause lower error than some classic method such as Pearson correlation coefficient.Then,an algorithm based on records of views from multiple users together is introduced to solve the sparsity problem of users' grades on services.Finally,this thesis explains the design and implementation of Mashup platform.The platform is developed based on an open source project called NODE-RED.The main works are packaging REST services into the form of the platform service and deploying service matching and recommendation modules.