INFORM: A Tool for Classification and Semantic Annotation of IoT Datastreams

Internet of Things (IoT) is transforming every aspect of our life and driving innovation in nearly every application domain areas including Smart Cities. An important aspect of IoT is in its ability to generate fine grained data and, in particular, the ability to understand and contextualize such data in order to (re)use it effectively. However, the metadata (a description of data produced by IoT) required to make sense of IoT datastreams is mostly inaccurate, incomplete and, in many situations, only human interpretable. Lack of metadata diminishes the value and effective use of IoT datastreams. To address this challenge, we propose INFORM, a tool to automatically classify, annotate and semantically enrich IoT datastreams. INFORM incorporates a novel technique to 1) accurately classify IoT datastreams using both the numerical characteristics of the data and limited available metadata, and 2) enrich the IoT datastreams with domain specific metadata extracted from application domain ontologies such as air pollution and sensor ontologies such as IoT-Lite. This paper presents a proof of concept implementation of the proposed INFORM tool as well as empirical evaluations to validate the effectiveness and efficiency of the datastream annotation process with INFORM.