Trade-Offs in a Google Distance and a WordNet Hybrid for QoS-Enabled Web Services Composition

Abstract

This paper proposes a hybrid approach in using Google Distance and WordNet together in a new method for the semantic similarity matching stage of web services discovery. We provide comparisons, using services recall and precision metrics, between our hybrid approach and our earlier lightweight Google Distance-based approach for web services discovery. Our performance evaluation demonstrates the trade-offs between the Google Distance only approach and a WordNet-assisted Google Distance hybrid approach for similarity matching in the web services discovery phase. Further, the impact of both approaches on QoS-enabled web service composition is described for 7 representative web transactions in the Travel domain.