Non-salient region erasure for time series augmentation  

作　　者：Pin LIU Xiaohui GUO Bin SHI Rui WANG Tianyu WO Xudong LIU 

机构地区：[1]State Key Lab of Software Development Environment(SKLSDE),Beihang University,Beijing,100191,China [2]Beijing Advanced Innovation Center for Big Data and Brain Computing(BDBC),Beijing,100191,China [3]Hangzhou Innovation Institute,Beihang University,Hangzhou,310052,China [4]School of Computer Science and Technology,Xi’an Jiaotong University,Xi’an,710049,China

出　　处：《Frontiers of Computer Science》2022年第6期203-205,共3页中国计算机科学前沿（英文版）

基　　金：supported by the National Key Research and Development Program (2018YFB1306000);Ministry of Industry and Information Technology of China (2105-370171-07-02-860873);State Key Lab of Software Development Environment (SKLSDE);Beijing Advanced Innovation Center for Big Data and Brain Computing (BDBC).

摘　　要：1 Introduction Time seriesaugmentationis an essential approachto solvethe overfitting problem on the time series classification(TSC)task[1,2].Although existing approaches perform better in mitigating this problem,none of them focus on protecting saliency regions on time series.The key informative shapelets contained in these regions are the core basis for distinguishing categories(e.g.,upward spikes in ECG and high amplitude in Sensor).

关 键 词：SENSOR SERIES SERIES 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]