基于频谱注意力机制和编码解码模型的时间序列分类研究.pdfVIP

Abstract

Timeseriesreferstoreal—valuedseriesdatawhoseattributeshaveorderexists

relationship．It

inoursociallifeandfield．Theoftraditionaltimeseriesclassifica—

widelyengineeringapplication

tionoftentheiscumbersomeandnotuniversal．

algorithmsrequiresexpertexperience，SOalgorithm

Themodeliscanlearntheinternalofthe

end—to-end，which

learningautomatically

deep