• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference in Advances in Electrical and Computer Technologies >Performance Analysis of Machine Learning Algorithms for IoT-Based Human Activity Recognition
【24h】

Performance Analysis of Machine Learning Algorithms for IoT-Based Human Activity Recognition

机译：基于IOT的人类活动识别机器学习算法性能分析

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Enormous growth is seen in the field of information technology and communication technology which enable the Internet of Things (IoT) technology on the boon. Nowadays, IoT technology is very common and widely used in all areas. It is also expanding its wing day by day. At the end of 2019, IoT devices will reach to 26.66 billion. Due to wide and diverse use of IoT technology, a large amount of valuable data is generated which gives researchers to deal with the huge amounts of real-time data. Machine learning plays a vital role in making a IoT environment, but it is a very complex task to build a smart environment. For personal healthcare monitoring, IoT-based sensors and mobiles devices play a crucial role in the betterment of the human lifestyle. Wearable sensor technology which is incorporated with mobile devices is more commonly used in monitoring personal health and well-being. In this research work, we examine the capability and performance of machine learning algorithms over the built-human activity recognition (HAR) dataset. Based on the performance evaluation results, gradient boosting classifier (GBC), support vector machine (SVM), random forest (RF), bagging classifier (BAG), classification and regression trees (CART), k-nearest neighbors (KNN), and extra trees classifier (ETC) algorithms have the better accuracy and suitable for real IoT datasets.
机译：在信息技术和通信技术领域看到巨大的增长，使得福音上的东西（物联网）技术能够实现。如今，物联网技术非常普遍，广泛应用于所有领域。它也在一天中扩展其翼。 2019年底，物联网设备将达到266.6亿。由于IOT技术的广泛和多样化，因此产生了大量的有价值的数据，使研究人员能够处理大量的实时数据。机器学习在制作IOT环境方面发挥着至关重要的作用，但它是一个非常复杂的任务来构建智能环境。对于个人医疗保健监控，基于物联网的传感器和移动设备在提高人类生活方式方面发挥着至关重要的作用。与移动设备合并的可穿戴传感器技术更常用于监测个人健康和福祉。在本研究工作中，我们研究了在内置人类活动识别（HAR）数据集中的机器学习算法的能力和性能。根据性能评价的结果，梯度升压分类器（GBC），支持向量机（SVM），随机森林（RF），装袋分类器（BAG），分类和回归树（CART），k-最近邻（KNN），和额外的树木分类器（ETC）算法具有更好的准确性，适合真正的IOT数据集。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号