Page 44 - 《水资源与水工程学报》2025年第1期
P. 44
!36 "!1 # & ' ( ) & * + , - Vol.36No.1
2025 $ 2 % JournalofWaterResources&WaterEngineering Feb.,2025
DOI:10.11705/j.issn.1672-643X.2025.01.05
's Reg-Crossformer fOW#ä4µ¦
,GH, IJK
( E9a0 í00G, ¡Á FU 211100)
> ?: Ö³EJKAB@A, ^+´¯Z[ 1961—2015 T{âÿJíê, eL£ç Crossformer iº, b
FÚ'p´, 'ÚÜëÏÎ@<íê, ¢Ý4©ªÒë~|Î{so。#¯®/Ü^5úQÞ{
Reg-Crossformer iº, ðñ³EJKÿJ, Æ}Éóàuiº+,Z[òó>Ã{È'。®«S¤:Reg-
Crossformer iº6¨ Crossformer iº, Istí(R) ² {Ñtí(NSE) µTé¯ 7.46% 21.63%, 3
"Ð6(RMSE) ÊÓ¯ 15.25%; ¦Òiº6Ãq4,Reg-Crossformer iº¦Á(³´ü¬3»2p
0iº(SVM) \æ~0iº(LSTM Informer), S0#¢»{ij{«í·Ñ。Reg-Crossformer iº
ìàðñ³EJKJ¯}.,{, ôÎJKZ]^\æ~0iº¦Z[ij}KÃr
q1´î。
1@A: ÿJÏÎ@<ðñ; æ~0;Reg-Crossformer; ³EJK
BC<DE:P333 FGHIJ:A FKLE:1672643X(2025)01004007
DailyrunoffpredictionbasedonReg-Crossformermodel
WANGKunyu,LIUXiangyang
(SchoolofMathematics,HohaiUniversity,Nanjing211100,China)
Abstract:TakingtheWeiheRiverBasinastheresearchbackground,thedailyrunoffdataofXianyang
stationfrom1961to2015areselectedasthedatainputfortheCrossformermodel.Focusingonthemulti
dimensionaltimeseriesdata ,thispaperadoptsthetwostageattention(TSA)mechanismtobettercapture
thecorrelationbetweendifferentdimensions.Inaddition,theReg-Crossformermodelincorporating
multisourcecovariatesisproposedtofurtherenhancetheadaptabilityofthemodeltocomplexhydrological
conditions.TheresultsofdailyrunoffpredictionintheWeiheRiverBasinshowthatcomparedwiththeo
riginalCrossformermodel ,theproposedmodelimprovesthecorrelationcoefficient(R)andNashefficien
cycoefficient (NSE)by7.46% and21.63% respectively;reducestherootmeansquareerror(RMSE)
by15.25%.Inthecomparativeexperimentsofdifferentmodels,Reg-Crossformeroutperformsthecon
ventionalmachinelearningmodel (SVM)anddeeplearningmodels(LSTMandInformer)acrossvarious
evaluationindicators,demonstratingsuperiorsimulationperformanceandstability.Reg-Crossformer
modeloffersanewapproachfortheaccuratepredictionofrunoffintheWeiheRiverBasin,andprovides
valuableinsightsintofutureapplicationofwaterresourcesmanagementanddeeplearningmodelsinhydro
logicalsimulation.
Keywords:dailyrunofftimeseriesprediction;deeplearning;Reg-Crossformer;theWeiheRiverBasin
m?W [1] 。xï, ðñøùú¼、 ¼
1 ./01
ÆDNv$%{-89, JKtút$Î$
JKZ[ðñZ]^{sîJª: +,, IZ[LDrµñÏ1fgÑ、 Ü^úûå
µ, p0Z]^¡¢b、 þSl /:c>{;6 [24] 。Z[ðñ}K{;6VW/
!"#$:20240305; %$:20240818
'()*: !"wx#0%»?@(3(41830110)
+,-.: )²g(1999—), Æ, ¡ÁFUN, ÈPABS, VWXYÏÎ@<ðñÔ{AB
qr+,: 5Ô¯(1976—), K, Ç~8ûN, OP, QR, ÈTSTU, VWXYíêµ¶、 p0、 áȸ¹:r
{AB。