A framework of VI-based ranking and recommendation of web services

There are several QoS-based approaches for web service selection. Every approach may use different QoS model to rank web services, which make clients confused. In reality, users want to choose most used web services with better trade-off between quality and cost. So referring to Alexa, this paper proposes a framework for collecting QoS and cost of web services and counting visits and introduces VI (Visits Index) to indicate the popularity of web services and argue VI. VI not only reflects total visits, but also reflects the rate of visits increment. Users can select web services based on VI ranking. For web service providers, how to balance properties of QoS and cost to make as many visits as possible is a problem. This paper uses ANN to establish the relation of QoS and cost to VI. Web service providers can use the trained ANN published by the framework to predict VI ranking of their web services and design non-functional properties to promote visits. The framework can be used for open web service marketplaces to rank web services and analyze correlation between non-functional properties and visits of web services similar to Alexa for websites.