Be4SeD: benchmarking for evaluation of service discovery techniques

Abstract

The growing number of business processes and services, has resulted countless develop research projects, aimed at generating discovery mechanisms. While there are a number of algorithms to retrieve services, such work does not use a common basis to evaluate the used search techniques, preventing objective evaluations. Therefore, it is necessary a public tool that provides a common reference for comparing and evaluating the results of the algorithms used in the matching of services, in order to improve the quality of the proposed discovery techniques. This article presents a public application, which implementing a benchmarking methodology to evaluate the quality of recovery of service matching techniques. This benchmarking is comprised of an intuitive assessment mechanism, an input module data for the algorithm to evaluate and a component that provides statistical results: recall, precision, overall, k-precision and p-precision. Its features are offered as a web service to facilitate integration with the implementations of algorithms to evaluate. Finally a matching algorithm is evaluated for evidence the use of Be4SeD platform in this context.