Averange-based fuzzy time series merkov chain untuk meramalkan kurs nilai tukar USD-IDR

Nur, Wahyuni (2015) Averange-based fuzzy time series merkov chain untuk meramalkan kurs nilai tukar USD-IDR. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
5211100095-Undergraduate_Thesis.pdf

Download (3MB)

Abstract

Perubahan nilai tukar mata uang sangat berpengaruh terhadap aktivitas ekonomi seperti investasi, struktur produksi perusahaan, perdangangan internasional (ekspor & impor), serta berpengaruh pada harga pasar produk & gaji pekerja. Pergerakan atau fluktuasi nilai tukar yang susah diprediksi menyebabkan investor, importir & eksportir mengalami kesulitan dalam memutuskan untuk melakukan transaksi bisnis. Fuzzy time series adalah salah satu model yang biasa digunakan dalam memprediksi data time series. Pada tugas akhir ini, untuk meramalkan nilai kurs valuta asing USD-IDR akan menggunakan model Average based Fuzzy time series Markov chain. Model ini merupakan gabungan dari 4 konsep yaitu: konsep fuzzy, konsep time series, konsep markov chain dan juga konsep average based. Konsep fuzzy digunakan untuk mengklasifikasi variabel, konsep time series digunakan untuk mengobservasi data nilai kurs pada periode waktu tertentu, konsep average base digunakan untuk menentukan interval efektif dan untuk menentukan transisi matriks probabilitas dalam proses peramalan menggunakan konsep markov chain. Output dari tugas akhir ini adalah peramalan nilai kurs valuta asing pada satu hari kerja berikutnya. Pada tugas akhir ini, hasil peramalan dengan average based fuzzy time series markov chain akan dibandingkan dengan hasil dari metode peramalan lain yaitu fuzzy time series dan moving average. Tingkat keakurasian akan diukur menggunakan MAPE. Hasil dan temuan dari penelitian ini adalah model average-based fuzzy time series markov chain memiliki peforma peramalan harian yang sangat baik dalam meramalkan kurs nilai tukar USD-IDR. Hal ini ditunjukkan dengan uji coba perbandingan nilai peramalan yang menggunakan 1230 data harian dari bulan Januari 2010 sampai dengan Desember 2014 yang menghasilkan nilai MAPE sebesar 0.33%. Dibandingkan dengan metode peramalan lain yaitu Fuzzy Time Series dan moving average, model peramalan average-based fuzzy time series markov chain mampu menghasilkan peramalan yang lebih baik dalam meramalkan kurs nilai tukar USD-IDR. Dilihat dari nilai MAPE pengujian. =================================================================================================== Changes of exchange rates influence on economic activity such as investment, production structure of the company, the international trading (export and import), as well as the effect on the market price of the product and salary workers. The exchange rate movements are difficult to predict, and it cause investors, importers and exporters had difficulty in deciding to conduct business transactions. The fuzzy time series is one of method that usually applied to forecast time series data. In this study, for forecast exchange rate between USD-IDR will use Average-based Fuzzy Time Series Markov chain model. This model is combination from 4 concept, that is fuzzy concept, time series, markov chain and average based consept. Fuzzy consept used to classify variables, the concept of time series of data is used to observe the exchange rate at a specific time period, the concept of average-based used to determine the effective interval and for determine the transition probability matrix in the forecasting process using the concept of Markov chain. The output of this study is forecasting the value of exchange rate on the next business day. In this study, the forecasting results with average based fuzzy time series Markov chain will be compared with the results of other forecasting methods (fuzzy time series and moving average). Accuracy rate will be measured using MAPE. The Result of this research is average-based fuzzy time series Markov chain model has an excellent performance in daily forecasting of exchange rate USD-IDR. This is demonstrated by testing using the comparative value forecasting 1230 daily data from January 2010 to December 2014 which resulted in the value of MAPE is 0.33%. Compared with other forecasting methods, Fuzzy Time Series and moving averages, average-based fuzzy time series Markov chain model is able to produce better result in forecasting the exchange rate USD-IDR. Based from the value of MAPE forecast.

Item Type: Thesis (Undergraduate)
Additional Information: RSSI 519.233 Wah a
Uncontrolled Keywords: Average-based; Fuzzy Time Series; MAPE; Markov chain; Moving average;Peramalan; Valuta asing
Subjects: Q Science > QA Mathematics > QA274.7 Markov processes--Mathematical models.
Divisions: Faculty of Information and Communication Technology > Information Systems > (S1) Undergraduate Theses
Depositing User: - Taufiq Rahmanu
Date Deposited: 30 Sep 2019 02:59
Last Modified: 30 Sep 2019 02:59
URI: http://repository.its.ac.id/id/eprint/70912

Actions (login required)

View Item View Item