Machine learning

PatchTST

PatchTST je arhitektura zasnovana na peščevima (patch-based Transformer) za prognoziranje vremenskih serija, koju su uveli Nie i saradnici 2023. godine, a koja seče svaku seriju na preklapajuće peščeve tretirane kao tokeni i obrađuje kanale nezavisno. Ona balansira računsku efikasnost sa visokom preciznošću na dugoročnim prognozama.

Otvorite u MethodMindUskoroVideoUskoroDownload slides

Pročitajte celu metodu

Samo za članove

Prijavite se besplatnim nalogom da biste pročitali ovaj odeljak.

Prijavite se

Method map

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

+4 more

Izvori

  1. Nie, Y., Nguyen, N. H., Sinthong, P. & Kalagnanam, J. (2023). A Time Series is Worth 64 Words: Long-term Forecasting with Transformers. ICLR. link
  2. Zhou, T., Ma, Z., Wen, Q., Wang, X., Sun, L. & Jin, R. (2022). FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting. ICML. link

Kako citirati ovu stranicu

ScholarGate. (2026, June 1). Patch Time Series Transformer. ScholarGate. https://scholargate.app/sr/deep-learning/patchtst

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

Citirana u

ScholarGatePatchTST (Patch Time Series Transformer). Preuzeto 2026-06-15 sa https://scholargate.app/sr/deep-learning/patchtst · Skup podataka: https://doi.org/10.5281/zenodo.20539026