基于可穿戴式纳米生物传感器的人体运动数据挖掘算法

Data mining algorithm for human motion based on wearable nano biosensor

针对当前人体运动数据挖掘算法无法对实时数据进行采集与分析,导致人体运动数据挖掘正确率较低且时间较长的问题,提出基于可穿戴式纳米生物传感器的人体运动数据挖掘算法。首先,利用可穿戴式纳米生物传感器采集人体运动数据,将采集到的数据转换为二进制数据形式,并对转换后的数据进行清洗与补位处理;最后,使用萤火虫算法对K均值聚类方法进行优化,利用优化后的K均值聚类方法对清洗与补位后的数据进行聚类处理。实验结果表明,所提算法的召回率平均值为97.12%,数据挖掘正确率平均值为98.42%,为运动员生理指标的实时监测与分析提供重要的数据基础。

To address the problem that current human motion data mining algorithms cannot collect and analyze real-time data,which leads to low correct rate and long time for human motion data mining,we propose a human motion data mining algorithm based on wearable nano-biosensor.Firstly,the human motion data is collected using the wearable nano-biosensor.Secondly,the collected data are converted into binary data form,and the converted data are cleaned and complemented.Finally,the K-mean clustering method is optimized using the firefly algorithm,and the optimized K-mean clustering method is used to cluster the cleaned and patched data.The results show that the average recall rate of the proposed algorithm is 97.12%and the average correct data mining rate is 98.42%,which provides an important data base for real-time monitoring and analysis of athletes′physiological indexes.