Peramalan pupuk ZA menggunakan Threshold Generalized Autoregressive Conditional Heteroskedasticity (TGARCH) (studi kasus di PT Petrokimia Gresik) / Hayu Aprilia
| Main Author: | Aprilia, Hayu |
|---|---|
| Format: | Thesis NonPeerReviewed |
| Terbitan: |
, 2017
|
| Subjects: | |
| Online Access: |
http://repository.um.ac.id/17498/ |
Daftar Isi:
- ABSTRAKApriliaHayu.2017.PeramalanPupukZAMenggunakanThresholdGeneralizedAutoregressiveConditionalHeteroscedasticity(TGARCH).SkripsiJurusanMatematikaFakultasMatematikadanIlmuPengetahuanAlamUniversitasNegeriMalang.PembimbingTrianingsihEniLS.SiM.Si.KatakunciperamalanheteroskedastisitasasimetrisTGARCH.Halutamayangdilihatdariprosesperamalanadalahdataberasumsihomoskedastisitas.Namunhaliniseringtidakterpenuhiyaituadanyaheteroskedastisitasdimanavariansidatatidakkonstan.ModelyangbisadigunakanuntukmengatasikasusheteroskedastisitasadalahmodelARCHGARCH.ModelGARCHmemenuhiasumsiguncangannegatifdanpositifyangberpengaruhsamaterhadapnilaikeruncingannyahalinitidakdapatmengatasiefekasimetris.TerjadinyaefekasimetrismengakibatkanmodelARCHGARCHtidaktepatuntukditerapkandalammeramalkandata.DalammengatasihalinisalahsatumetodeyangdapatdigunakanadalahperluasandarimodelGARCHyaituTGARCH.ModelTGARCHdiperkenalkanolehZakoaian(1994).ModelThresholdGeneralizedAutoregressiveConditionalHeteroscedasticity(TGARCH)inimerupakansalahsatumodelkasusheteroskedastisitas.ThresholdiniberupavariabeldummyyangditambahkanpadamodelGARCHdenganmaksuduntukmengakomodirkemungkinanterjadinyaasimetrisdalamvolatilitassuatuvariabelsebagaiakibatbadnewsdangoodnews.ModelTGARCHterbaikdalammeramalkanpupukZAyangdikirimkeGudangPenyanggaBukittinggiSumateraBaratpadaperiodeJanuari2012sampaiMei2016yaituARIMA(110)-TGARCH(11)dengansatuthresholddenganpersamaanyaituZ_t(1552533)Z_(t-1)-(0552533)Z_(t-2)a_tdengan12310963_t123112-0956197N_(t-1)945_(t-1)20570059963_(t-1)2.HasilperamalanpupukZAkeBukittinggiSumateraBaratpada2periodeselanjutnyayaituperiodeJunidanJuli2016sebanyak1779832tonpupukdan2128828tonpupuksedangkandaridataaslisebanyak200tonpupukdanbulanJuli2016tidakadapengirimanpupukdengannilaiakurasipadagrafikstaticdidapatkannilaiBiasProportion(BP)sebesar00026dannilaiVarianceProportion(VP)sebesar082sedangkanpadagrafikdynamicdidapatkannilaiBiasProportion(BP)sebesar00036nilaiVarianceProportion(VP)sebesar0984.NilaiBiasProportion(BP)mendekatinoldanVarianceProportion(VP)mendekati1sehingganilairamalansudahdikatakanmendekatinilaisebenarnyadengandemikianmodelARIMA(110)-TGARCH(11)dengan1thresholdmerupakanmodelyangbaikuntukmeramalkandataalihstokkeBukittinggiPTPetrokimiaGresik.ABSTRACTApriliaHayu.2017.ForecastingofZAfertilizerusingThresholdGeneralizedAutoregressiveConditionalHeteroscedasticity(TGARCH).ThesisMajoringofMathematicFacultyofMathematicandSainsStateUniversityofMalang.AdviserTrianingsihEniLS.SiM.Si.KeywordsforecastinghetroskedasticityasymetrisTGARCH.Themainthingseenfromtheforecastingprocessistheassumedhomoscedasticitydata.Butthisisoftennotfulfilledietheheteroskedastisitaswherethevarianceofdataisnotconstant.ThemodelthatcanbeusedtoovercomecasesofheteroskedastisitasisARCHGARCHmodel.TheGARCHmodelmeetstheassumptionsofnegativeandpositiveshocksthathavethesameeffectontheircrestvaluethiscannotovercometheasymmetriceffect.TheoccurrenceofasymmetriceffectsresultedinanARCHGARCHmodelnotappropriatetobeappliedinpredictingdata.ToovercomethisoneofthemethodsthatcanbeusedistheextensionoftheGARCHmodelTGARCH.TheTGARCHmodelwasintroducedbyZakoaian(1994).TheThresholdGeneralizedAutoregressiveConditionalHeteroscedasticity(TGARCH)modelisoneofthecasesofheteroscedasticity.ThisthresholdisadummyvariableaddedtotheGARCHmodelinordertoaccommodatethepossibilityofasymmetryinthevolatilityofavariableasaresultofbadnewsandgoodnews.ThebestTGARCHmodelinforecastingZAfertilizersenttoBukittinggiWestSumatraWarehouseonJanuary2012untilMay2016isARIMA(110)-TGARCH(11)withonethresholdwithequationZ_t(1552533)Z_(t-1)-(0552533)Z_(t-2)a_twith12310963_t123112-0956197N_(t-1)945_(t-1)20570059963_(t-1)2.ResultsofforecastingofZAfertilizertoBukittinggiWestSumatrainthenexttwoperiodsofJuneandJuly2016asmuchas177.9832tonsoffertilizerand212.8828tonsoffertilizerwhilefromtheoriginaldataof200tonsoffertilizerandJuly2016thereisnofertilizerdeliverywithaccuracyvalueInthestaticgraphgotthevalueofBiasProportion(BP)equelto00026andthevalueofVarianceProportion(VP)equalto082whileindynamicgraphgotthevalueofBiasProportion(BP)00036andthevalueofVarianceProportion(VP)equelto0984.TheProportionBias(BP)valueisclosetozeroandVarianceProportion(VP)approaches1sothepredictedvalueissaidtobeclosetothetruevaluethustheARIMAmodel(110)-TGARCH(11)with1thresholdisamodelbothtoforecastthetransferofstockdatatoBukittinggiPTPetrokimiaGresik.1stExaminer2ndExaminerTrianingsihEniLestariS.SiM.SiJamaliatulBadriahS.PdM.SiNIP198301012005012001NIP198812302015042001MainExaminerIr.HendroPermadiM.SiNIP196612241999031001