A New Web Services Filtrating Mechanism with Latent Semantic Analysis

This paper analyzes some existing web services description models and services matching methods, puts forward a Web services filtrating mechanism based on latent semantic analysis to solve lower matching efficiency and precision of the available services discovery mechanisms that is short of services filtrating mechanism and matched with keywords. Firstly, this filtrating algorithm describes basic attributes and QoS attributes of Web services using tree structure attributes template. Secondly, the paper builds latent semantic analysis space using certain terms-frequency statistic method and weights method. And then, the paper builds advertising services index database, and filtrates the Web services according to the service request. Experiments results show that this algorithm has higher precision and recall ratio, can improve services matching efficiency largely.