運用時間數列分解法建構股價預測模型之研究

The Study of the Stock Price Prediction Model by Using Time Series Decomposition Analysis

陳信良(Hsin-Liang Chen)

指導教授 ： 古永嘉

國立臺北大學/商學院/企業管理學系/碩士(2012年)

摘要

本研究目的在探討上市電子股股價淡旺季指數特性，並進行投資模擬分析。研究對象以上市大盤、電子類及2010年底市值前50大電子股(2000年已上市)，自CMoney系統取2000年至2010年11年之132筆月股價資料。本研究運用SAS-X11模組進行季節穩定性及移動性測試，探討季節指數變化特性，並以月股價季節指數算出模擬投資報酬率。本研究實證結果如下：
1. 季節指數穩定性及移動性測試中，發現大盤及電子類均具有顯著差異。個股方面季節指數穩定性測試有42家具有顯著差異，顯示有42檔股票股價在一年當中，具顯著淡旺季差異；移動性測試則顯示50家均具有顯著差異，顯示在過去的11年當中，所有樣本股價其淡旺季指數有移動跡象。
2. 模擬投資報酬率，個股平均年報酬率，有46家優於大盤及電子類，有3家僅優於大盤，也有1家遜於大盤及電子類。而50檔股票平均年報酬率，發現僅2008年度遜於大盤及電子類，其餘年度則較優。

關鍵字

時間數列分解法 ； 電子產業 ； 季節指數穩定性及移動性測試 ； 股價預測模型

並列摘要

The study focuses on the characteristics of seasonal indices: carrying out simulated investment analysis by using stock price data of listed companies from electronic sector. It includes all TWSE listed companies, electronic sector companies and the top 50 largest market value (by year end of 2010) electronic sector companies. The monthly share price data (132 numbers of data in total) are obtained from CMoney system within the time frame: 2000 to 2011. SAS software (X11 module) is employed to perform seasonal stability and movement ability test. And then, we analyzed the characteristics of the changes in seasonal indices and calculated the simulated investment rate of return by using monthly stock prices in seasonal indices.
Results of this empirical study:

1. By applying the seasonal stability and movement ability test, the difference between all listed companies and electronic sector companies is obviously noticeable. Among the stock prices of individual companies group, the seasonal stability test shows that there are 42 companies which have noticeable difference. This result shows the stock prices of these 42 companies vary significantly among high seasons and low seasons within the same year. Movement ability test indicates all top 50 companies have major price variations. This result shows that for the past 11 years, all selected stock price samples have a movement trail during high seasons and low seasons.
2. Under the simulated investment rate of return test, the average annual rate of return for 46 individual companies surpasses the performance of all listed companies and electronic sector companies. 3 of them are slightly better than that of all listed companies and electronic sector companies. One of them is worse than that of all listed companies and electronic sector companies. On the other hand, the average rate of return for the top 50 electronic sector companies is only worse than that of all listed companies and electronic sector companies in year 2008. For the rest of other years, their performances are better.

並列關鍵字

Time Series Decomposition Analysis ； Electronic sector companies ； seasonal stability and movement ability test ； Stock price prediction model

參考文獻

10. 黎明淵、林修葳、郭憲章、楊聲勇(民91)，美、日股市巨幅波動下的股市連動效果－美國、日本與亞洲四小龍股市實證結果，證券市場發展季刊，第57 期，pp. 117-143。

12. 鍾健能(民99)台灣食品業營收與營運資金供需季節指數關聯性之研究，國立台北大學企業管理學系碩士論文。

1. 王紹明(民100)，製藥產業營收與股價之季節指數特性與因果關係之實證研究，國立台北大學企業管理學系碩士論文。

1. Bamber, L. S. and Y. S. Cheon, (1995) “Differential Price and Volume Reactions to AccountingEarnings Announcements,” The Accounting Review, 62, pp.510-532.

2. Basic, E., S. Ozyildirim, and K. Aydogan, 1996 “A note on Price-Volume Dynamics in an Emerging Stock Market,” Journal of Banking and Finance, 20, pp.389-400.