台灣加權股價指數之預測模型:小波轉換與多項式迴歸模型之應用

A Prediction Model for Taiwan Stock Exchange Index: Application of the Wavelet Transform and Polynomial Regression

曾琬婷(Wan-Ting Tseng)

指導教授 ： 李顯峰

國立臺灣大學/社會科學院/經濟學研究所/碩士(2017年)

https://doi.org/10.6342/NTU201704306

摘要

本文研究台灣加權股價指數(TAIEX)的預測模型，因欲與三個預測模型比較預測的準確度，選擇1998-2006年為樣本期間，應用小波轉換(Wavelet Transform)方法進行比較。小波轉換克服傅立葉轉換(Fourier Transform)之不足之處，可以分析非定態(non-stationary)與非週期性資料，並同時凸顯時域與頻域之效果。

小波轉換擷取資料特徵值，並根據特徵值重構多項式迴歸混合預測模型。佐以網格搜索(Grid Search)方式，探索台灣加權股價指數金融時間序列資料的趨勢及變化。實證研究結果顯示，本研究所應用的小波轉換與多項式迴歸預測模型可準確地預測台灣加權股價指數，其預測結果較其他現行研究結果的預測模型更為優異，具有參考價值，可提供決策主管當局之參考。

關鍵字

小波轉換 ； 台灣加權股價指數 ； 網格搜索 ； 多項式迴歸

並列摘要

In this study we try to construct a forecasting model of Taiwan Stock Exchange Index (TAIEX), based upon the Wavelet Transform method, to investigate the predicted accuracy from 1998 to 2006. Wavelet transform overcomes the shortcomings of Fourier transform, i.e., it can analyze non-stationary and non-periodic time series data, and highlights the effect of time domain and frequency domain.

Wavelet transform extracts the eigenvalue of the sample data and combines with the polynomial regression according to the eigenvalue. Grid Search explores the trend and change of financial time series data of Taiwan Stock Exchange Index. Our major findings show Wavelet transform combined with polynomial regression model can more accurately predict the Taiwan Stock Exchange Index. The predicting power of our model is more competent than the three previous models. It can be of a better reference for by the decision-making authorities.

並列關鍵字

Wavelet Transform ； Taiwan Stock Exchange Index (TAIEX) ； Grid Search ； polynomial regression

參考文獻

Aizerman, M.-A., Braverman, E.-M. and Rozonoer, L.-I., 1964, “Theoretical Foundations of the Potential Function Method in Pattern Recognition Learning,” Autom. Remote Control, Vol.25, 821-837.

Mousazadeh, Abbasi N., Aghaei, M. A. and Moradzadeh, Fard, M., 2015, “Forecasting Stock Market Using Wavelet Transforms And Neural Networks And ARIMA (Case Study Of Price Index Of Tehran Stock Exchange),” International Journal of Applied Operational Research, Vol.5, No.3, 31-40.

Bernstein, P.-L., 2009, Capital Ideas Evolving, USA: Wiley.

Bagheri, Ahmad, Hamed Mohammadi Peyhani and Mohsen Akbari, 2014, “Financial Forecasting using ANFIS Networks with Quantum-behaved Particle Swarm Optimization,” Expert Systems With Applications, Vol.41, No.14, 6235-6250.

Cheng, Shou-Hsiung, Chen, Shyi-Ming and Jian, Wen-Shan Jian, 2016, “Fuzzy Time Series Forecasting Based on Fuzzy Logical,” Information Sciences,Vol.327, 272-287.