ARIMA

PERAMALAN PENJUALAN PRODUK MINUMAN “X” AREA PEMASARAN JAWA TIMUR DENGAN MENGGUNAKAN METODE VARIMA

“FORECASTING THE SALE OF “”X”” BEVERAGE PRODUCT IN EAST JAVA REGION OF MARKETING BY USING VARIMA METHOD”

Created by :
Kurniawan, C. Ade ( )

Subject: Box-Jenkins forecasting
Alt. Subject : Box-Jenkins forecasting
Keyword: Varima
sales forecasting

[ Description ]

Penjualan produk minuman “X” pada area yang saling berdekatan diduga terdapat korelasi atau hubungan saling mempengaruhi. Dalam penelitian ini akan digunakan pendekatan deret waktu multivariat, yaitu pemodelan Vector Autoregressive Integrated Moving Average (VARIMA). Penelitian dilakukan terhadap 4 kelompok area sebagai vektor deret waktu yang dibentuk dari 9 area pemasaran di Jawa Timur berdasarkan rekomendasi perusahaan dan nilai korelasi antar daerah pemasaran. Setelah dilakukan pemodelan dengan menggunakan program paket SAS, berdasarkan nilai AIC didapatkan orde AR dan MA yang tinggi untuk masing-masing vektor deret waktu. Akan tetapi, pemodelan tidak dapat dilakukan karena keterbatasan kemampuan iterasi konvergensi optimasi parameter pada program paket SAS. Selain itu, terdapat outlier, sehingga berdasarkan ketentuan dalam metode VARIMA, model yang dihasilkan dalam penelitian ini be-lum optimal. Hal itu diketahui dari asumsi distibusi normal multivariat dan white noise dalam vektor residual yang tidak terpenuhi. Oleh karena itu, dilakukan pemodelan secara univariat dengan menggunakan metode ARIMA. Berdasarkan kriteria out sampel (MAPE dan MAD), model ARIMA lebih baik (kecuali area Lamongan). Hasil Pemodelan dengan metode ARIMA pada semua area, tidak mengandung model AR dan dipengaruhi data outlier. Hal itu menunjukkan, penjualan produk “X” memiliki kecenderungan stabil, tidak ada tren naik, namun sering muncul outlier.

Alt. Description

According to the approximation that there are correlation among the sale of “X” beverage product at area which nearby each other, this research will use multivariat time series approach. That is Vector Autoregressive Integrated Moving Average (VARIMA). Research conducte to 4 group as vector time series which is formed from 9 marketing areas in East Java that pursuant to company’s recommendation and the correlation value of interregional marketing. After the modelling process conducted by using SAS, pursuant to AIC, the models of each vector time series contains high order of AR and MA. However,the modelling process can’t be conducted because of the limitation of ability to convergence the iteration of parameter optimization of SAS. Others, there are outliers, so that pursuant to provisions in method VARIMA, model yielded in this research isn’t optimal. That matter is known from normal multivariate distribution assumption and white noise process in the residual vector which is not fullfiled. Therefore, univariate modeling by using ARIMA method is conducted. Its result that according to out sample criterion (MAPE and MAD), ARIMA model (except Lamongan) is better than VARIMA model. The result of ARIMA modelling has no AR models with influence from outliers. It show that, sale of “X” beverage product owning stable tendency, no tren yet, but often influenced by outliers.

Contributor :
  1. Ir. Dwiatmono Agus Widodo, M.Ikom
Date Create : 23/12/2009
Type : Text
Format : pdf
Language : Indonesian
Identifier : ITS-Undergraduate-3100008032663
Collection ID : 3100008032663
Call Number : RSSt 519.535 Kur p

Source :
Undergraduate Theses of Department of Statistics Faculty of Mathematics and Natural Science RSSt 519.535 Kur p, 2006

Coverage :
ITS Community Only

Rights :
Copyright @2008 by ITS Library. This p ublication is protected by copyright and permission should be obtained from the ITS Library prior to any prohibited reproduction, storage in a retrievel system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to ITS Library