基于混沌分形理論的我國期貨市場波動性研究.pdf

1、青島大學碩士學位論文基于混沌分形理論的我國期貨市場波動性研究姓名：王霞申請學位級別：碩士專業(yè)：金融學（含：保險學）指導教師：伍海華20060515AnalysisofFluctuationsofChineseFutureMarketBasedonChaosandFraetalTheoryAbstractStartingfromthemodernfinancialfrontiercomplexsystemwelookedbackbasic

2、approachesandprinciplesofthechaosfractaltheoryandFractalMarketHypothesisWeshowedthatthechaosandfractaltheorycouldreflecttlleessenceofthefinancialmarketascomplexsystemBasedonFractalMarketHypothesis，westudiedthebehaviorsof

3、thefuturemarketinthemethodsoffractaltheorythatwasonebranchofcomplexsystem0urdiscussionfocusedonthecharacteristicsoffractalsystem：longtermcorrelationandstatisticalsimilarityWeshowedthatthetwopropertieswereconsistentasintr

4、insicpropertyofcomplexsystemIntheempiricalanalysis，wefoundstrongevidenceoflong—termpersistenceandstatisticalsimilarityonChinesefuturemarketUsingR／Sanalysis，weestimatedHurstExponenttocharacterizethelongtermpersistenceTheH

5、urstExponentsestimatedwerestablymorethan05inalltimescalesanditmeantthattherewaslong—termpersistenceinChinesefuturemarketnamelytllebehaviorsofthefuturemarketfollowedtheskewrandomwalkmodelInR／Sanalysiswefound血eevidenceof‘N

6、on—periodicCycle”infuturemarketwithaveragecyclebetween75daysand80daysInthetermsoffractalstatisticsfractalmarketalsoexistedAtierthechangesoftimescalesthescattersandtheprobabilitydistributionsalsemainedthesimilarshapes，whi

7、lemisshowedtherewerethestructuresofsimilarityinthefuturemarketAlsobyanalyzingtheLyapunovexponentsoffuturepricesseries，weestimatedthepredictedtimeaslongaspossiblewitllthelongestsome78daysAnditshowedthatⅡleempiricalresults

8、ofchaosandfractaltheorywereconsistentinⅡleerrorscaleInsummarytheempiricalresultsdemonstratedthatthefuturemarketwasofintrinsicpropertyofchaosandfractalsystemWefurtherconductedElliott昭1償analysisofthechaoticpredictionandapp