Data Stream Learning in IoT Environment: A Survey

Abstract

Smart systems need to take faster decisions. Computation of stored data induces delays and renders a real time system useless. Sensors are not smart devices, they collect data at edge level of network. Uploading the data to cloud for processing is slow, a faster way to predict from the data collected is by processing it at the lower levels of network. Machine Learning concepts can be exploited to train the collaborating devices to represent such data for information exchange. This paper discusses background of possible research areas open for proposing decision making systems at lower layers of network hierarchy