Optimization of CART Decision Tree Algorithm in Tiny Deep Learning System to Improve IOT Sustainability

Optimization of CART Decision Tree Algorithm in Tiny Deep Learning System to Improve IOT Sustainability

—Tiny deep learning is deployed on local or edge IoT devices instead of processing in the cloud, and uses machine learning by embedding artificial intelligence in the hardware. One of the growing areas of deep learning is small and is a subset of machine learning, algorithms, hardware and software. It aims to enable low-latency interfaces in edge devices that consume only a few milliwatts of battery power. Such reductions in consumption enable machine learning devices to last for weeks, months, or even years while the machine learning application is running continuously at the edge or endpoint. In this article, we introduce a Tiny deep learning system using the CART decision tree and describe the applications of this system in Internet of Things networks. Machine learning also solves data security, privacy, latency, storage and energy efficiency issues. Considering the use of decision trees in the layers of neural networks in this article, the performance of these networks has increased, and this performance, along with small deep learning algorithms, has reduced network load and energy consumption in Internet of Things environments.