ScholarGate
Assistent
Machine learningTime-series forecasting

LightTS: Let sampling-orienteret MLP til multivariat tidsserieprognose

LightTS er en letvægts, MLP-baseret arkitektur til multivariat tidsserieprognose, introduceret af Tianping Zhang og kolleger i 2022. Motiveret af observationen, at simplere modeller kan matche eller overgå tunge Transformer-baserede arkitekturer, anvender LightTS en interval-samplingstrategi til at dekomponere lange inputsekvenser i flere undersekvenser og behandler hver med kompakte Chunk-MLP- og Continuous-MLP-moduler. Designet prioriterer beregningseffektivitet, samtidig med at det bevarer både lokale og globale tidsmæssige mønstre.

Åbn i MethodMindSnartVideoSnartDownload slides

Læs hele metoden

Kun for medlemmer

Log ind med en gratis konto for at læse dette afsnit.

Log ind

Method map

The neighbourhood of related methods — select a node to explore.

LightTS: Let sampling-orienteret MLP til multivariat tidsserieprognose
DLinear: Decomposition L…Multilayer Perceptron (M…TSMixer: All-MLP Arkitek…

Kilder

  1. Zhang, T., Zhang, Y., Cao, W., Bian, J., Yi, X., Zheng, S., & Li, J. (2022). Less is more: Fast multivariate time series forecasting with light sampling-oriented MLP structures. arXiv preprint. link

Sådan citerer du denne side

ScholarGate. (2026, June 2). LightTS (Light Sampling-oriented MLP). ScholarGate. https://scholargate.app/da/deep-learning/lightts

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side
ScholarGateLightTS (LightTS (Light Sampling-oriented MLP)). Hentet 2026-06-15 fra https://scholargate.app/da/deep-learning/lightts · Datasæt: https://doi.org/10.5281/zenodo.20539026