WEB SERVICES RECOMMENDATION LEVERAGING SEMANTIC SIMILARITY COMPUTING

With the popularity of Web services adopted for supporting domain applications, recommending and composing appropriate services with respect to user requirements is a challenge. This paper proposes a dynamic programming and variable length genetic algorithm for the recommendation and composition of Web services. Generally, starting and ending services are determined leveraging the constructed service network model. Based on which, services are selected and composed, such that these services should be more appropriate on satisfying users' requirements. Experimental evaluation result shows that our technique is effective and can improve the accuracy of service recommendation.