基于局部距离特征的多模态融合CNN时间序列分类  

MULTI-MODAL FUSION CNN TIME SERIES CLASSIFICATION BASED ON LOCAL DISTANCE FEATURE

作　　者：马志强[1,2] 石磊 Ma Zhiqiang;Shi Lei(Henan Vocational College of Agriculture,Zhengzhou 451450,Henan,China;School of Information Engineering,Zhengzhou University,Zhengzhou 450001,Henan,China)

机构地区：[1]河南农业职业学院,河南郑州451450 [2]郑州大学信息工程学院,河南郑州450001

出　　处：《计算机应用与软件》2023年第4期241-250,共10页Computer Applications and Software

基　　金：国家重点研发计划(科技部)项目(2018YFB1701400)。

摘　　要：针对动态时间规整时间序列分类中存在的动态匹配信息丢失问题,提出一种基于局部距离特征的多模态融合CNN时间序列分类。提出一种新的特征,称为局部距离特征,当给定DTW(Dynamic Time Warping)的非对称斜率约束时,即使对于不同长度的时间序列,局部距离特征向量的长度均可以固定,并且增加了动态时间规整输入对于时序内在特征失真的鲁棒性。进一步在多模态融合CNN中结合了局部距离特征和坐标特征。探索不同的原型选择方法、原型数目和数据融合方案对精度的影响。通过对典型的时间序列数据集进行分类实验,实验结果表明该方法可以有效提高分类精度。Aimed at the problem of missing dynamic matching information in dynamic time warping time series classification,a multi-modal fusion CNN time series classification based on local distance feature is proposed.A new feature called local distance feature was proposed.When the asymmetric slope constraint of dynamic time warping(DTW)was given,the length of local distance feature vector could be fixed even for time series of different lengths.The robustness of dynamic time warping input to the intrinsic feature distortion of time series was added.The local distance feature and coordinate feature were combined in the multi-modal fusion CNN.The effects of different prototype selection methods,prototype numbers and data fusion schemes on the accuracy were explored.The classification experiments results on typical time series datasets show that the proposed method can effectively improve the classification accuracy.

关 键 词：卷积神经网络 时间序列分类 动态时间规整 距离特征 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP3[自动化与计算机技术—控制科学与工程]