Trust Assessment on Streaming Data: A Real Time Predictive Approach

Tao Peng, Sana Sellami, Omar Boucelma

Abstract: IoT data, that most often carry a temporal dimension, can be exploited from an analysis perspective or from a forecasting one. In this paper, we propose a predictive approach to address the problem of data trustworthiness in a data stream generated by a Smart Home application. We describe an online Ensemble Regression model that performs prediction in assigning a trust score to a target temporal value in real-time. Experiments conducted with data retrieved from the UCI ML repository demonstrate the performance of the model, while assessing data accuracy.