PERAMALAN EKSPOR PERIKANAN INDONESIA DENGAN PENDEKATAN ARIMA, FEED FORWARD NEURAL NETWORK, DAN WEIGHTED FUZZY TIME SERIES

PARAMESWARI, EUNIKE WIDYA (2016) PERAMALAN EKSPOR PERIKANAN INDONESIA DENGAN PENDEKATAN ARIMA, FEED FORWARD NEURAL NETWORK, DAN WEIGHTED FUZZY TIME SERIES. Masters thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 1314201706-Abstract.pdf]
Preview
Text
1314201706-Abstract.pdf - Published Version

Download (321kB) | Preview
[thumbnail of 1314201706-Master Thesis.pdf]
Preview
Text
1314201706-Master Thesis.pdf - Published Version

Download (3MB) | Preview
[thumbnail of 1314201706-conclusion.pdf]
Preview
Text
1314201706-conclusion.pdf - Published Version

Download (261kB) | Preview

Abstract

Salah satu indikator untuk memonitor peningkatan perdagangan internasional adalah
dengan memperhatikan pertumbuhan ekspor. Peningkatan kinerja ekspor mempunyai
dampak langsung dalam meningkatkan pertumbuhan ekonomi. Hal ini berarti ekspor
memiliki dampak langsung terhadap perekonomian suatu negara. Ketersediaan data
ekspor sangat penting bagi pengambilan keputusan dan kebijakan perdagangan luar
negeri oleh pemerintah. Berdasarkan klasifikasi komoditi, salah satu sektor yang
memberikan kontribusi cukup besar terhadap volume ekspor Indonesia adalah
perikanan. Ketersediaan data volume ekspor terkini menjadi tantangan dalam
mengaplikasikan metode peramalan yang efektif. Pada awalnya, metode peramalan
didominasi oleh metode linier. Namun demikian, metode linier tidak dapat menangkap
hubungan non-linier yang seringkali dijumpai pada kondisi riil. Tujuan dari penelitian
ini adalah untuk meramalkan volume ekspor perikanan Indonesia dengan metode linier
serta non-linier. Metode linier yang digunakan adalah ARIMA, sedangkan metode nonlinier
yang digunakan adalah Feed Forward Neural Network (FFNN) dan Weighted
Fuzzy Time Series (WFTS). FFNN merupakan arsitektur NN yang sering diaplikasikan
dalam berbagai bidang. Metode ARIMA digunakan untuk mendapatkan arsitektur NN
yang paling sesuai sehingga dapat diperoleh model NN dengan kinerja peramalan
terbaik. Hasil empiris dari penelitian ini menunjukkan bahwa metode WFTS unggul
dalam peramalan pada kelompok komoditi 302 dan 306. Sedangkan metode Hibrida
ARIMA-NN menunjukkan kinerja yang terbaik pada peramalan untuk kelompok
komoditi 303.

=====================================================================================================

One of the indicators to monitor the increase in international trade is by paying
attention to the export growth. Improved export performance has a direct impact in
promoting economic growth. This means that exports have a direct impact on the
economy of a country. Export data availability is critical for decision-making and
foreign trade policy by the government. Based on the classification of commodities, one
of the sectors that make a significant contribution to the Indonesian export volume is
fisheries. Current export volume of data availability is a challenge in applying an
effective forecasting method. At first, the method of forecasting is dominated by linear
methods. However, the linear method can not capture non-linear relationships that are
often found in real conditions. The aim of this study is to predict the volume of
Indonesian fisheries exports to the method of linear and non-linear. Linear methods
used are ARIMA, whereas non-linear method used is Feed Forward Neural Network
(FFNN) and Weighted Fuzzy Time Series (WFTS). FFNN is an NN architecture that is
often applied in various fields. ARIMA method used to obtain the most suitable NN
architecture so as to obtain NN models with the best forecasting performance. The
empirical results of this study indicate that the method WFTS outperforms in
forecasting the commodity groups of 302 and 306. While hybrid ARIMA-NN method
showed the best performance in forecasting for 303 commodity groups.

Item Type: Thesis (Masters)
Additional Information: RTSt 519.535 Par p
Uncontrolled Keywords: ekspor, Neural Network, Weighted Fuzzy Time Series , ARIMA
Subjects: Q Science > QA Mathematics
Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression
Divisions: Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis
Depositing User: Mr. Tondo Indra Nyata
Date Deposited: 11 Jan 2017 06:55
Last Modified: 27 Dec 2018 03:31
URI: http://repository.its.ac.id/id/eprint/1482

Actions (login required)

View Item View Item