Domain Agnostic Quality of Information Metrics in IoT-Based Smart Environments
(From the article abstract)
Thanks to the proliferation of IoT devices that are interconnected, huge amounts of data are being gathered nowadays. The availability of all these new sensors, data sources and open data platforms offers new possibilities for innovative applications and use-cases that are many times dynamic. However, if we plan to depend on data for the optimal provision of services, it is of utmost importance to ensure the quality of data and the quality of information that we are handling in an online manner.
Furthermore, geolocalised data provides a richer context in which the quality of information can be measured and in which services are more advanced. In order to support the process of finding the right information, we have defined several metrics in single-sensor and multi-sensor scenarios that are based on statistical analysis, machine learning algorithms and contextual information. We have applied them in two scenarios: smart parking and environmental sensing for smart buildings.
In many studies, complex algorithms claim to preserve data quality, however, they lack straightforward definitions of what quality is. In this work, we define metrics
for DQ and compute them in several IoT scenarios for checking their viability.