Capmcm24201197_亚太杯数学建模竞赛一等奖.docxVIP

Team#apmcTeamNumber:

apmcProblemChosen:

C

Researchonanalysisandcountermeasuresofpetindustrydevelopmentbasedonintegratedmachinelearning

Summary

ThisarticleexaminesthedevelopmentstatusandtrendsofthepetindustryinChinaandglobally.Thearticlecombinesdatafromthepastfiveyearstoconductresearchfromfourperspectives:thecurrentstatusandforecastoftheChinesepetindustry,theforecastofglobalpetfooddemand,ananalysisofChinesepetfoodproductionandexports,andtheimpactofinternationaltradepolicies.Usingcorrelationanalysis,multiplemachinelearningmodels,andfusiontechnology,thearticleanalyseskeyissuesandproposescorrespondingcountermeasurestoprovidescientificguidancefortheindustrysfuturedevelopment.

Fortask1,TheobjectiveofthisstudyistoforecastthedevelopmentoftheChinesepetindustryoverthenextthreeyears.Inthedatapreprocessingstage,thispaperestablishedanindicatorsystemforthedevelopmentofthepetindustryandrelatedindustries,correctedmissingvaluesandoutliers,anduseddescriptivestatistics,Box-Coxtransformation,andfeaturescalingtooptimisethedatastructuretoensuredataquality.TheChinesepetmarketispredictedforthenextthreeyearsbasedonmachinelearningmodels,includingGBDT,XGBoost,andRUSBoost,andthroughtheStackingfusionmodel.Thefindingsindicateagradualincreaseinthenumberofpetcats,withanestimated79.85millionin2024.Thenumberofpetdogsexhibitsslightfluctuations,yettheoverallgrowthtrendpersists,reaching51.25millionin2024.TheresultsofthemodelverificationdemonstratethattheStackingmodelexhibitsthehighestgoodnessoffit(R2reaching0.986),renderingitanoptimalchoiceforforecastingtheprospectivetrajectoryofChinaspetindustry.

Fortask2,Aforecastofglobaldemandforpetfoodispresentedherewith.Amachinelearningmodelwasconstructe