摘要
维度的升高会加剧运动行为传感数据的复杂度,导致其分布特征空间被无限放大,因此提出基于有限维空间的运动行为传感数据特征提取方法。采用关联规则项挖掘分析方法计算数据模糊度,确定运动行为的有限空间区域。在有限维空间下,通过自适应寻优方法,计算传感数据的特征量化参数。检测运动行为传感数据的特征属性,计算数据分布融合映射输出结果,构建运动行为特征提取模型。实验结果表明,所提方法的运动数据空间聚类效果较好,能够把数据固定在有限维空间,数据特征提取精度始终保持在95%以上。
The increase of dimension will increase the complexity of physical behavior sensing data,resulting in its distribution feature space being infinitely enlarged.Therefore,a feature extraction method for physical behavior sensing data based on finite dimensional space is proposed.The association rule mining analysis method is used to calculate the data ambiguity and determine the limited space region of the physical behavior.In finite dimensional space,the eigenvalues of sensing data are calculated by adaptive optimization method.The feature attributes of physical behavior sensing data are detected,the output of data distribution fusion mapping is calculated,and the motion behavior feature extraction model is constructed.The experiment results show that the proposed method has a good spatial clustering effect on physical data,can fix the data in a finite dimensional space,and the accuracy of data feature extraction is always above 95%.
作者
卢瑛
LU Ying(Xi’an Siyuan University,Xi’an 710038,China)
出处
《信息技术》
2024年第7期115-120,共6页
Information Technology
基金
陕西省体育局课题(2021095)。
关键词
有限维空间
运动行为
传感数据
关联规则项挖掘
特征提取
finite dimensional space
physical behavior
sensing data
mining association rule items
feature extraction
