Research on IoT Edge Data Mining Techniques Based on Cloud Computing Models

Author:

Kang Yang1,Li Qiang1,Liu Yuanyong1

Affiliation:

1. Chongqing University of Arts and Sciences , Yongchuan , Chongqing , , China .

Abstract

Abstract With the continuous development of the Internet of Things (IoT) and the needs of the times, the number of IoT edge devices is rapidly increasing, and to be able to provide users with more reliable information services and decision-making, this paper constructs an IoT system based on cloud-edge collaboration. Based on cloud-edge collaboration, data mining technology is used to categorize data, and regression models and branch neural networks are combined to further improve the IoT system based on cloud-edge collaboration. Finally, the particle swarm algorithm is used to optimize reinforcement learning, and the framework of an IoT system built on cloud-edge partnership is established. Through the data processing experiments on IoT edge computing, the data mining performance of the IoT edge computing-based data mining method is improved by 0.0295, 0.1870, and 0.2059 on the large datasets of Dataset4 and Dataset5, respectively. the average turnaround time is lower than 6.5 for different edge computing power levels. The system rewards are gradually improved from 5.3 to 8.2, and the number of packet loss of the algorithm is less than 5. The feasibility and effectiveness of the IoT data mining method and system based on cloud-edge collaboration are confirmed by this study, and a feasible method for data mining can be applied to IoT systems under cloud-edge collaboration.

Publisher

Walter de Gruyter GmbH

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3