ARIMA model-based web services trustworthiness evaluation and prediction

Meng Li,Zhebang Hua,Junfeng Zhao,Yanzhen Zou,Bing Xie
DOI: https://doi.org/10.1007/978-3-642-34321-6_51
2012-01-01
Abstract:As most Web services are delivered by third parties over unreliable Internet and are late bound at run-time, it is reasonable and useful to evaluate and predict the trustworthiness of Web services. In this paper, we propose an ARIMA model-based approach to evaluate and predict Web services trustworthiness. First, we evaluate Web services trustworthiness with comprehensive trustworthy evidences collected from the Internet on a regular basis. Then, the cumulative trustworthiness evaluation records are modeled as time series. Finally, we propose an ARIMA model-based multi-step Web services trustworthiness prediction process, which can automatically and iteratively identify and optimize the model to fit the trustworthiness series data. Experiments conducted on a large-scale real-world data set show that our method can effectively evaluate and predict the trustworthiness of Web services, which helps users to reuse Web services.
What problem does this paper attempt to address?