INTERNET OF THINGS EDGE DATA MINING TECHNOLOGY BASED ON CLOUD COMPUTING MODEL

. Due to the limited computing resources of Internet of Things edge devices, cloud and edge computing can work together to improve the way data is processed and services are delivered. Cloud computing focuses on providing centralized resources, while edge computing pushes data processing to the edge of the network, reducing latency. To this end, the research proposes to use edge computing for data processing on the cloud computing platform to reduce the pressure of network bandwidth. It utilizes the addition of early exit branch neural networks to process data on edge devices. Simultaneously through a linear regression model, the optimal branch exit point is determined based on network latency and device load for enhancing the efficiency and accuracy of data classification. The results showed that the training accuracy of AlexNet and Resnet networks with branches also increased with the increase of training times. The training accuracy of Layer 2 was maintained within the range of 0.98-1.0 and 0.99-1.0, respectively. When the latency was 1.25 ms, the classification accuracy of the branch AlexNet network differed the most from that of the unbranched AlexNet network by 15%. When the delay was 3.5 ms, the classification accuracy of the branch Resnet network was 4.1% higher than that of the unbranched Resnet network. The classification inference performance of the research model is the best, with a performance detection value of 0.92. The research method can serve as a new model for addressing the requirements of the Internet of Things and localized computing, and has important reference value in reducing the latency and data transmission bandwidth of computing systems.