Update Artikelilmiah: 13416

REZA NUR HIDAYATULLOH

NIM

Judul Artikel

Abstrak (Bhs. Indonesia)

Tujuandaripenelitianiniadalahuntukmemahamipemodelan ARIMA danpenerapanmetode ARIMA dalamperamalanhargaemaspadabulanJanuari 2015 – Juni 2015.Dalampenelitianinidijelaskantentanglangkah-langkahdalampembentukan model ARIMA. Tahapawaldaripemodelan ARIMA adalahidentifikasistasioneritas data, denganCorrelogramfungsiautokorelasi(ACFdan PACF).Identifikasibentuk ACF dan PACF dari data yang sudahstasionerdigunakanuntukmenentukanorde model ARIMA dugaan.Tahapanselanjutnyaadalahestimasi parameter dancekdiagnosauntukmelihatkesesuaian model. Proses cek diagnose dilakukanuntukmengevaluasiapakah residual model sudahmemenuhisyarat white noise danberdistribusi normal. Unit Root Testadalahuji yang dapatdigunakanuntukmemvalidasisyaratwhite noise, sedangkanUjiNormalitas Residualmerupakanujiuntukevaluasidistribusi normal. Hasilujinormalitas residual menunjukkanbahwa residual model ARIMA sudah white noise.
Hasil penelitian inimenunjukkan model yang digunakanuntukmemprediksihargaemasadalahmodel ARIMA (4,1,4) tanpakonstantadenganpersamaanYt =  (0,788) Yt-4 – (-0,935) et-4 + et.Hasilperamalanmenunjukkanbahwa model ARIMA inidapatdigunakanuntukmemprediksihargaemas, dibuktikandenganhasilpengukurankesalahanperamalan MAD 38,69, MSE 3395,73, dan MAPE 3,23%. Hasil MAPE sebesar 3,23% menunjukkanbahwa model relevanuntukdigunakandalamperamalan.

Abtrak (Bhs. Inggris)

The purpose of this research was to understand the ARIMA modeling and application of methods ARIMA forecasting gold prices in January 2015 - June 2015. In this research, explained about the steps in the formation of the ARIMA model. The early stages of stationary ARIMA modeling is the identification of the data, with Correlogram autocorrelation function (ACF and PACF). Identification forms of ACF and PACF of the data that has been used to determine the order stationary ARIMA models allegations. The next stage is the estimation of the parameters and diagnostics check to see the suitability of the model. The process of diagnosis checks performed to evaluate whether residual models already qualified white noise and normally distributed. Test Unit Root Test is a test that can be used to validate the condition of white noise, whereas Residual Normality Test is a test for the evaluation of the normal distribution. The test results show that the residuals normality residual ARIMA model has white noise.
To prove the hypothesis, this research using ARIMA (4,1,4) without constant. Model results showed ARIMA (4,1,4) without constants can be used to predict the price of gold. Evidenced by the results of measurements of forecasting error MAD 38.69, 3395.73 MSE and MAPE 3.23%. 3.23% MAPE results show that the model is relevant for use in forecasting.

Keywords: Gold Price, ARIMA.

Kata kunci

Pembimbing 1

Pembimbing 2

Pembimbing 3

Tahun

Jumlah Halaman