Machine Learning applied in the Edge-Devices

Internet of Things (IoT) has been of great interest during the recent decades. IoT network is a combination of millions of devices establishing a massive hierarchy. Exchanging data between the layers in the dimensions of thousand kilobytes per seconds. In a network, the devices doing computations at the lowest hierarchy are said to be the edge- devices. Their task is to receive, the data from the sensor, store it and then move it forward to the upper layer for computation and feature analysis. The edge devices, push the dataset without ensuring that they can have some sort of missing values or incorrect instances. It forwards the dataset for computation into the main server, leads to centralized computing. The unwanted datasets travel unnecessarily through the network, consuming plenty of bandwidth of the network. Devices that are present at the network edge can generate the way of decentralized processing. The approach is to shift some computation from the main server to the edge devices. The key is to add the decision-making capabilities on the Edge-devices. This can be achieved by using a classification learning algorithm.The result stated that applying machine learning, decision tree, assures that edge-devices can be made reliable for classification with an accuracy of up to 98 percentages. Hence, taking a step closer to the path of learning for fast-computing.