資源描述：
《Time Series Forecasting》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、TimeSeriesForecastingMatiasG.EnzTimeSeriesForecasting?Quickreviewofforecasting?DifferentApproachestoForecasting?Introductiontotimeseries?MeasuringForecastError?Describingatimeseries?ForecastingmethodsforleveldemandWhatisForecasting??Forecastingisthepredi
2、ction,projectionorestimationoftheoccurrencesofuncertainfutureeventsorlevelsofactivity.?Examples:?Weatherthisweekend?Stockprices?Nextmonth’ssalesinregionX?…?ResultsofpoorforecastingFundamentalForecastingApproachesWhatdoesthefuturehold??Qualitative?Relatio
3、nal?TimeSeriesTimeSeriesApproaches?Consistsofdatathatarecollected,recorded,orobservedoversuccessiveincrementsoftime?Assumespastbehaviorisgoingtoberepeatedinthefuture?Searchforpatternsinthedata--decompositionintodifferentcomponents?Manytimeseriestechnique
4、sDescribingaTimeSeries?Wecanthinkofahistoricaltimeseriesasbeingcomposedof“pattern”and“noise.”?the“pattern”canincludestationarity,trend,seasonality,oracycle,?the“noise”israndomwithzeromeanandseriallyuncorrelated?Thegoalofforecastingistoforecastthepattern.
5、?Noiseisrandomandcan’tbeforecasted.ComponentsoftheTimeSeriesOriginaltimeseriesDecomposedtimeseriesCyclecycleSeasonalseasonalityDemandDemandTrendtrendLevellevelTimeNoisenoiseTimeDemandcanbeexpressedasafunctionofthesecomponents,forexample,D(t)=(L+Tt)[S(t)]
6、+rChartstakenfrom“OperationsManagement,”R.G.Schroeder,IrwinMcGraw-Hill,2000ForecastOptimality?Aforecastisoptimalif:?allpatternhasbeendiscovered?onlyerrorcomesfromthenoiseterm?Forecastfitreferstohowtheforecastmatches-upagainstpastdata?over-fitting(forecas
7、tingtherandomness)isbadForecastError?Ameaningfulforecastconsistsof2values:?Pointestimate+error?Forecasterrorinformationisusedfor:?Selectingforecastingmethods?Estimatingsafetystocks?Improvingforecastingprocess?EvaluatingriskSomeTimeSeriesTechniques?Simple
8、?Complex?Movingaverages?Spectralanalysis?Exponential?Box-Jenkinssmoothing(ARIMA)?Decomposition?KalmanFilters?FourierAnalysis?EvaluatetheforecastusingnewPastdataisavailable.demanddataasitcomesin....DDDDDDD...t-3t-2t-1tt+1t+
