Leveraging Machine Learning in IoT to Predict the Trustworthiness of Mobile Crowd Sensing Data

Abstract

The advances in Internet-of-things (IoT) have fostered the development of new technologies to sense and monitor the urban scenarios. Specifically, Mobile Crowd Sensing (MCS) represents one of the suitable solutions because it easily enables the integration of smartphones collecting massive ubiquitous data at relatively low cost. However, MCS can be affected by wrong data-collection procedures by non-expert practitioners, which can be make useless (or even counter-productive), if contributed data are not trustworthy. Contextualizing monitored data with those coming from phone-embedded sensors and from time/space proximity can improve data trustworthiness. This work focuses on the development of a machine learning method that exploits context awareness to improve the reliability of MCS collected data. It has been validated on a case study concerning urban noise pollution data and promises to improve the trustworthiness of data generated by end users.