Model Statistik Arima Dalam Meramal Pergerakan Harga Saham
Abstract
Abstrak: Berbagai teori dan analisa dikembangkan untuk antisipasi baik risiko tak sistematik maupun risiko sistematik. Teori portofolio yang dikemukakan oleh Harry Markowitz menekankan cara diversifikasi atas investasi untuk mengurangi jenis risiko tak sistematik dan mengoptimalkan tingkat keuntungan berdasarkan analisa fundamental. Sedangkan analisis teknikal yang menyangkut ekonometrika model Autoregressive ARIMA (p,d,q) (demikian pula model NN, neural network) dikembangkan untuk mengantisipasi risiko pasar secara umum (sistematik dan tak sistematik), dengan berdasarkan pada data historis harga saham itu sendiri atau disebut data time series, dengan asumsi investor bertindak logis dalam mengambil keputusan investasi dan bersikap menghindari risiko investasi, membawa konsekuensi pada trend pergerakan harga yang berbentuk pola tertentu atau setidaknya trend yang terbentuk akan bertahan cukup lama. Dengan kata lain tidak mengikuti pola random walk atau perilaku spekulatif yang menyebabkan harga saham bersifat fluktuatif. Berdasarkan hasil uji pemodelan terhadap 53 saham yang tergabung dalam LQ45 periode 2005-2006, ternyata hanya 19 saham yang memiliki model peramalan kuantitatif ARIMA.Kata kunci: ARIMA, LQ45, random walk, stasionir, time series.
Abstract: Various theories and analyzes were developed to anticipate both non-systematic risk and systematic risk. Portfolio theory put forward by Harry Markowitz emphasizes ways to diversify investment to reduce the types of non-systematic risk and optimize the level of profit based on fundamental analysis. While technical analysis concerning econometrics of the ARIMA Autoregressive model (p, d, q) (as well as the NN model, neural network) was developed to anticipate general market risk (systematic and non-systematic), based on historical data of the stock price itself or referred to time series data, assuming investors act logically in making investment decisions and avoid investment risks, have consequences for price movement trends that take the form of certain patterns or at least the formed trend will last long enough. In other words, it does not follow random walk patterns or speculative behavior that causes stock prices to fluctuate. Based on the results of the modeling test of 53 shares incorporated in LQ45 for the period 2005-2006, it turns out that only 19 shares have quantitative forecasting model of ARIMA.
Keywords: ARIMA, LQ45, random walk, stasionir,time series.
References
[1] Carol. A, Market Models, A Guide to Financial Data Analysis. UK : John Wiley & Sons Ltd, 2001.
[2] Edwin. J, Elton, Martin, J. Gruber, Stephen, J. Brown, Modern Portofolio Theory and Investment Analysis, sixth Edition, USA: Jhon Wiley & Sons.Inc, 2003.
[3] Daniel Pena, George, C. Tiao, Ruey, S. Tsay. A Course in Time Series Analysis. Canada. hon Wiley & Sons. Inc. 2001.
[4] Enders. W, Applied Econometric Time Series. New York.: John Wiley & Sons,1995.
[5] George, E.P.Box, Gwilym, M. Jenkins, Time Series Analysis, forecasting and control, third Edition. USA : Printice Hall. Inc,1994.
[6] James, D. Hamilton. Time Series Analysis. New Jersey : Princeton University Press,1994.
[7] Walter Enders, Applied Econometric Time Series, USA : Jhon Wiley & Sons. Inc, 1995.
[8] Frank, R. Giordano, Maurice, D. Weir. A First Course in Mathematical Modeling, USA : Brooks/Cole Publishing Company,1997.
[9] Koutsoyiannis. A. Theory of Econometrics, London & Basingstoke. : The Macmillan Press Ltd,1978.
[10] Katz, David. A. Econometric Theory and Applications, Englewood Cliffs. N.J.: Prentice Hall Inc,1982.