研究生: |
李俊賢 Li, Chun Hsien |
---|---|
論文名稱: |
利用複合式的最小平方法改善時間序列的多步預測問題 Improved Multi-step Forecasts in Time Series by Composite Least Square Methods |
指導教授: |
徐南蓉
Hsu, Nan Jung |
口試委員: |
蔡恆修
Tsai, Henghsiu 張雅梅 Chang, Ya Mei |
學位類別: |
碩士 Master |
系所名稱: |
理學院 - 統計學研究所 Institute of Statistics |
論文出版年: | 2016 |
畢業學年度: | 104 |
語文別: | 中文 |
論文頁數: | 90 |
中文關鍵詞: | recursive 方法 、direct 方法 、最佳線性預測 、最小平方法 、時間序列 、多步預測 、AIC 、結合預測 、leave-h-out cross-validation |
外文關鍵詞: | recursive method, direct method, the best linear prediction, least squares, time series, multi-step forecast, AIC, combined forecast, leave-h-out cross-validation |
相關次數: | 點閱:2 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
時間序列分析常用兩種方法進行多步預測,一種是透過重複代入一步最佳
線性預測的方式(recursive method) 遞迴地運算多步預測,另一種則是直接進行多步最佳線性預測(direct method)。兩方法所採用的參數估計量亦不相同,前者多採用MLE 或一步最小平方估計,後者則直接採用多步最小平方估計。相較而言,若模式能正確選取,recursive 方法所得到的多步預測均方差會較小;direct的方法參數估計誤差雖較大,但為不偏預測,且對模式的選取相對穩健。由於recursive 與direct 兩種預測方法互有優劣,本研究提出一個結合兩種預測方法之複合式的參數估計式,並透過資料訊息自動地調控估計式複合之權重,以期新方法能兼具recursive 與direct 預測的優點。本文將此創新的預測方法運用於多種線性與非線性的時間序列資料上,透過模擬研究比較傳統方法與新方法的預測能力優劣表現,並建議各種方法的最佳適用情境。
Multi-step forecast is an important issue in time series analysis. Among linear forecasts, the direct and recursive methods are both popular in use. The former
is solved by minimizing the h-step-ahead prediction mean squared error directly. The latter, also called plug-in or iterated method, is recursively computing the
multi-step-ahead prediction by repeatedly plug in the one-step-ahead best linear predictors to unobserved lag variables. Both methods are theoretically justified,
while their empirical performance relative to the other is depending on the tradeoff between the bias and estimation variance which is typically sensitive to the working model, the forecast horizon, and the underlying data scenario. This thesis proposes a composite inference for parameter estimation as well as for the multi-step
forecasting by combining the estimating functions from both direct and recursive methods. This new composite method is easy to compute and is expected to remain the advantages from both sides. In particular, the new method can automatically adjust the optimal weights between both predictors via a cross validation approach. Simulation studies show that the proposed composite forecast performs effectively and adaptive towards the better one among the traditional direct and recursive forecasts under a variety of linear and nonlinear data generating scenarios. Some practical recommendations and computational issues are also addressed.
[1] 張育瑋(2009), TAR 模型建模之相關議題, 清大統計所碩士論文
[2] 張景弦(2015), 局部線性外插與指數平滑法, 清大統計所碩士論文
[3] Chang, J.S., & Huang, L.S.(2015), Local Linear Extrapolation and Exponential Smoothing. working paper.
[4] Bhansali, R.J. (1997), Direct Autoregressive Predictors for Multistep Prediction: Order Selection and Performance Relative to The Plug in Predictors. Statistic Sinica, 7, 425-449.
[5] Chen, R.(2015), Functional Coefficient Autoregressive Models: Estimate and Tests of Hypotheses. working paper.
[6] Cheng, T.C.F., Ing, C.K., & Yu, S.H.(2015), Toward optimal model averaging in regression models with time series errors. Journal of
Econometrics, 189, 321-334.
[7] Findley, D.F.(1984), On Some Ambiguities Associated with The Fitting of ARMA Models to Time Series. Journal of Time Series Analysis, 5, 213-225.
[8] Hansen, B.E.(2008), Least-Squares Forecast Averaging. Journal of Econometrics, 146, 342-350.
[9] Hansen, B.E.(2010), Multi-Step Forecast Model Selection. working paper.
[10] Taieb, S.B., & Hyndman, R.J.(2012), Recursive and Direct Multi-step Forecasting: The Best of Both Worlds. working paper.
[11] Taieb, S.B., & Hyndman, R.J.(2014), Boosting Multi-step Autoregressive Forecasts. working paper.
[12] Hurvich, C.M., & Tsai, C.L.(1997), Selection of A Multistep Linear Predictor for Short Time Series. Statistica Sinica, 7, 395-406.
[13] Kunitomo, N., & Yamamoto, T.(1997), Properties of Predictors in Misspecified Autoregression Time Series Models. Jouranl of the American Statistical Association, 80, 941-950. 85
[14] Casella, G., Fienberg, S., & Olkin, I.(2014), An Introduction to Statistical Learning with Applications in R. New York, USA: Springer.
[15] Ing, C.K.(2003), Multistep Prediction in Autoregressive Processes. Econometric Theory, 19, 254-279.
[16] Yang, Y.(2003), Regression with Multiple Candidate Models: Selecting or Mixing ?. Statistica Sinica, 13, 783-809.