混沌時間序列預測研究及其應用.pdf

1、廈門大學碩士學位論文混沌時間序列預測研究及其應用姓名：王永忠申請學位級別：碩士專業(yè)：系統(tǒng)工程指導教師：曾昭磐19990301堡鎏墮塑壁!!堡!!查墊嬰塑!!AbstractInthispaperthestatespacepredictionmethodsforachaotictimeseriesbytimedelayreconstruction(alsoreferredtoasdelayreconstructionandphasespa

2、s#reconstruction)havebeenstudiedTheordinarychaotictimeseriespredicationmethodssuchas”zeroorderapproximation””largestLyapunovexponent”and’’locallinearregression”areresearchedInordertoreducecalculationerrors，anewmethodforp

3、redictingthechaotictimeseriesisproposedAnewforecastingmodelbasedonpointsofneighborsinthestatespaceandlinearregressiontechnologyissuggestedThepaperisorganizedasfollowsChapter1MajorfeaturesandparametersofchaosBasedonLorenz

4、andRosslermodel，weshowacharacteristicmeasureofstrangeattractorsandstudythefeatureofSDICFinallywegiveacalculationmethodofLyapunovexponentsandDimensionChapter2ChaotictimeseriesandtheirstimedelayreconstructionForchaoticthne

5、seriesfromLorenzandRosslermodel，westudythecorrelationdimension(D2)’embeddingdimension(m)，maximumofLyaponovexponents(x】)andpicturingtheirorbitofthereconstructedphasespace，Chapter3Introductiontopredictionmethodsofchaoticti

6、meseriesInthischaptertheordinarypredictionmethodssuchas”zero—orderapproximation”，”largestLyapunovexponent”and”locallinearregression”arcintroducedChapter4Researchsofchaotictimeseriespredictionmethods，Inthischapterwestudyt

7、hecauseofcalculationerrorsofordinaryChaotictimeseriespredicationmethodssuchas”zeroorderapproximation””largestLyapunovexponent”and’’locallinearregression”Aneweffectivepredictionmethodbasedonpointsofneighborsinthestatespac

8、ewithlinearregressiontechnologyisproposedTheeffectivenessoftheproposedmethodcomparingwithotherordinarymethodsisdemonstratedbythechaoticdatafromLorenzandRosslermodelChapter5Applicationsof’’LinearregressionbasedOilneighbor

9、spoints”methodTheeffectivenessofproposedmethodisexaminedbythefollowingpracticestimeseries：”sunspotannuajdata”，”Wuhanmonthlyaveragetemperaturedata!。，”Wuhanonearearatiodataofmonthlydeath“and”Xiamenharborsemi—monthlythrough