基于強(qiáng)化學(xué)習(xí)的績優(yōu)股票預(yù)測(cè)系統(tǒng)研究.pdf

1、國內(nèi)圖書分類號(hào)：TPl81國際圖書分類號(hào)：62L3工學(xué)碩士學(xué)位論文Y867325基于強(qiáng)化學(xué)習(xí)的績優(yōu)股票預(yù)測(cè)系統(tǒng)研究碩士研究生：導(dǎo)師：申請(qǐng)學(xué)位級(jí)別：學(xué)科、專業(yè)：所在單位：授予學(xué)位單位：楊櫻葉德謙教授工學(xué)碩上計(jì)算機(jī)應(yīng)用技術(shù)信息科學(xué)與工程學(xué)院燕山大學(xué)鲞當(dāng)盔蘭三蘭堡主蘭壘堡蘭AbstractAsestablishingnonlinearmodeusingartificialneuralnetworkhasbeenwidelyusedtoecon

2、omyforecastingforrecentlyyears，thispaperwilldosomeresearchonthetopicUsingBPneuralnetworksOnstockforecastingsystemisanewresearchfieldandhasagreatevolvementButtlliSmethodisnotgoodatstabilityanddefinitionSoitmakethesystem’S

3、selfstudyabilitynotstrong，whenthedataquantumbecomebigger，theforecastingresukbecomeweakerrefereacevalueInthispaperweprovidestheoptimizationonareinforcementlearningalgorithmbasedonneuralnetworkUsingthismethodwecanimproveth

4、esystem’Sstabilityanddefinitionandimprovethegeneralizationabilityoflearningsystem(1)Choosepropersocketdatatooperate，assurethedatahaveadefinitereferencevalue(2)Choosethepropermethodtopretreatmentthedata，assurethedatahaveg

5、oodastringencyandfasterconstringencyrateduringtheBPneuralnetworkstrainingprocess，thesystemperformancegaineddefiniteensure(3)Adopttheappropriatealgorithmbasedonneuralnetworkensemble，assuretheforecastingsystem’sstabilizati

6、onconquertheBPneuralnetworksystem’SinstabilitySothattheforecastingresulthavehigherreferencevalueandtheforecastingstockpricecalvehasbetterstability似)Analysisthecharacteristicoftimeserialadopttherightreinforcementlearninga

7、lgorithmtooptimizetheresultoftheneuralnetwork’SoutputchoosetheexcellentparameterofthereinforcementlearningsystemAllalgorithmsaboveareimplementedinMATLABandutilizingVCandMATLABrealizationthesystemconfirmexcellentparameter