PERAMALAN INDEKS HARGA SAHAM GABUNGAN (IHSG) DENGAN METODE FUZZY TIME SERIES MARKOV CHAIN

Y Aristyani(1), E Sugiharti(2),


(1) Jurusan Matematika, FMIPA, Universitas Negeri Semarang, Indonesia
(2) Jurusan Ilmu Komputer, FMIPA, Universitas Negeri Semarang, Indonesia

Abstract

Tujuan penelitian ini adalah untuk mengetahui akurasi metode Fuzzy Time Series Markov Chain pada peramalan IHSG dan membuat aplikasi untuk peramalan IHSG menggunakan software MATLAB. Dalam penelitian  ini, data bersumber dari yahoo finance. Data historis diambil dari data Composite Indeks (IHSG) periode Januari 2010 sampai dengan Februari 2014. Dengan mengubah data time series IHSG ke dalam fuzzy logic group untuk menentukan matriks probabilitas transisi, maka hasil peramalan dapat diperoleh. Tahap awal pembuatan aplikasi yaitu perancangan sistem. Aplikasi untuk peramalan IHSG dirancang dengan menggunakan GUI pada MATLAB dengan melakukan coding yang sesuai agar aplikasi bisa berjalan. Setelah dilakukan pengujian sistem diperoleh hasil MSE untuk metode Fuzzy Time Series Markov Chain sebesar 9827.1292 dan MSE untuk  metode Fuzzy Time Series S&C sebesar 15769.7036. Karena  memperoleh nilai MSE yang lebih kecil maka metode Fuzzy Time Series Markov Chain lebih akurat dan memiliki kinerja yang lebih baik untuk peramalan. Aplikasi yang dibuat memiliki persentase akurasi peramalan dengan metode Fuzzy Time Series Markov Chain sebesar 98,03458% dan persentase akurasi peramalan dengan metode Fuzzy Time Series S&C sebesar 97,38003%.

The purpose of  this research were to determine the accuracy of the Markov Chain Fuzzy Time Series method on JCI forecasting and make an application for JCI forecasting using MATLAB software. In this research, the data sourced from Yahoo Finance. Historical data is taken from Data Composite Index (JCI) in the period of January 2010 to February 2014. By transfering time series data into fuzzy logic groups to determine the transition probability matrix, then the forecasting results can be obtained. The initial phase to making the application  is system design. Application for JCI forecasting designed using GUI on MATLAB with appropriate coding in order to run the application. After testing the system then obtained MSE results for the Markov Chain Fuzzy Time Series method was 9827.1292 and MSE for Fuzzy Time Series S & C was15769.7036. Because the MSE values obtained was smaller then the method of Markov Chain Fuzzy Time Series is more accurate and has better performance for forecasting. Application are made have the percentage of forcasting accurary with Markov Chain Fuzzy Time Series method is 98,03458% and the percentage of forcasting accurary with Fuzzy Time Series S & C is 97,38003%.

Keywords

Fuzzy Time Series Markov Chain, IHSG

Full Text:

PDF

References

Arimbawa IBKP & Jayanegara K. 2013. Komparasi Metode ANFIS dan Fuzzy Time Series Kasus Peramalan Jumlah Wisatawan Australia Ke Bali. E-Jurnal Matematika 2 (2) 2013: 18-26.

Duru O & Yoshida S. 2009. Comparative Analysis of Fuzzy Time Series and Forecasting an Empirical Study of Forecasting Dry Bulk Shipping Index. Department of Maritime Transportation and Management Engineering, Istanbul Technical University. Turkey.

Hansun S. 2012. Peramalan Data IHSG Menggunakan Fuzzy Time Series. IJCCS 6 (2) 2012: 79-88.

Puspitasari I, Suparti & Wilandari Y. 2012. Analisis Indeks Harga Saham Gabungan (IHSG) Dengan Menggunakan Model Regresi Kernel. Jurnal Gaussian 1. Universitas Diponegoro.

Ross SM. 2003. Introduction to probability Models 10th Edition. University of Southern California. Los Angeles.

Tsaur RC. 2012. A Fuzzy Time Series Markov Chain Model With An Application to Forecast The Exchange Rate Between The Taiwan and US Dollar. International Journal of Innovative Computing, Information and Control. 8 7(B) 2012: 4931-4942

Refbacks

  • There are currently no refbacks.




Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.