基于深度学习的跳绳能耗测试仪错误数据检测方法  

Error data detection method of rope skipping energy consumption tester based on deep learning

作　　者：叶艳 YE Yan(School of Physical Education,Ankang University,Ankang Shanxi 725000,China)

机构地区：[1]安康学院体育学院,陕西安康725000

出　　处：《自动化与仪器仪表》2022年第8期277-280,283,共5页Automation & Instrumentation

基　　金：陕西省体育局常规课题《安康市体育旅游资源调查和旅游产业开发融合研究》(17104)。

摘　　要：针对跳绳能耗测试仪错误数据检测准确率低,错误数据漏报率、误报率高的问题,提出基于深度学习的跳绳能耗测试仪错误数据检测方法。使用小波变换技术完成原始数据预处理,为后续的错误数据检测提供数据基础。选用深度学习算法中的聚类分析方法,划分数据类别,完成数据分析。使用历史错误数据构建粗糙集,将数据聚类结果对其进行比对,确定错误数据集。至此,基于深度学习的跳绳能耗测试仪错误数据检测方法设计完成。实验结果表明:此方法数据检测准确率较高,错误数据漏报率、误报率偏低,实现了错误数据检测结果精度的整体提升。Aiming at the problems of low detection accuracy rate of error data of skipping rope energy consumption tester,false alarm rate and high false alarm rate,a deep learning-based error data detection method of skipping rope energy consumption tester is proposed.The original data preprocessing is completed by using wavelet transform technology,which provides the data basis for subsequent error data detection.The cluster analysis method in the deep learning algorithm to divide the data categories and complete the data analysis is used.A rough set is constructed using historical error data,and the data clustering results are compared to determine the error data set.So far,the design of the error data detection method for skipping rope energy consumption tester based on deep learning is completed.The experimental results show that the data detection accuracy of this method is high,the false data false negative rate and false positive rate are low,and the overall accuracy of false data detection results is improved.

关 键 词：异常数据检测 深度学习 跳绳能耗测试仪 回归模型 统计数据 残差分布 

分 类 号：TP306.3[自动化与计算机技术—计算机系统结构]