Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns
碩士 === 國立臺北大學 === 國際財務金融碩士在職專班 === 99 === According to Fama (1967, 1976), efficient market hypothesis is no longer fitted to the current stock market, therefore the stock prices can not reflect the real value of respective companies. Nevertheless, investors are able to gain excess returns through va...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2011
|
Online Access: | http://ndltd.ncl.edu.tw/handle/47176438969286790769 |
id |
ndltd-TW-099NTPU1304028 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-099NTPU13040282015-10-28T04:06:36Z http://ndltd.ncl.edu.tw/handle/47176438969286790769 Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns 以財務指標及技術指標建構股價預測模型-類神經網路模型之應用 Chen, Yen-Chen 陳鄢貞 碩士 國立臺北大學 國際財務金融碩士在職專班 99 According to Fama (1967, 1976), efficient market hypothesis is no longer fitted to the current stock market, therefore the stock prices can not reflect the real value of respective companies. Nevertheless, investors are able to gain excess returns through various approaches to predict stock prices. In reference to the article of Wyckoff (1970), the financial indices and technical indices are adopted as relative variances. The period of study span is from the year of 2000 to 2009. This study uses the consolidated financial data of underlying stocks of Taiwan 50 ETF and Taiwan 100 ETF as sample collection. Build up a forecasting model through the application of back-propagation network to analyze the effect of financial and technical indices and forecast the stock returns. Utilize predicted hit ratio (PHIT) to determine the validity of selected model and choose the best result for further study of investment simulation. Two investment strategies are developed. Strategy 1 is to buy the stocks which prices are predicted to go upwards. Strategy 2 is to sell the stocks which prices are estimated to decline in addition to buy the stocks which prices are predicted to go upwards. The study also compares with the model of fundamental indices (RAFI) on the validity and investment result. The empirical study shows: 1. Financial and technical indices have significant effects to stock returns. The R-square of stepwise regression has reached to 86.19%, it’s well matched. On the contrary, RAFI has no significant effect to stock returns. The R-square of stepwise regression only gets 5.37%. 2. The PHIT from the model of back-propagation network has reached to 90.91% and the result from RAFI also hits 67.33%. This reveals the model of artificial neural network has excellent performance to forecast stock returns. 3. The simulation result as follow, the investment combinations organized by strategy 1 and 2 have the average annual returns of 86.19% and 60.08% respectively; Sharp ratios are 1.20 and 0.96 as well. Meanwhile, RAFI gets the average annual return of 44.21% and 18.05% respectively; Sharp ratios are 0.61 and 0.21 separately. On the other hand, the average annual return in the market is 36.96% and sharp ratio is 0.47. In all, the performance from the model of back-propagation network is ahead of the results from RAFI and the market. Goo, Yeong-Jia 古永嘉 2011 學位論文 ; thesis 90 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立臺北大學 === 國際財務金融碩士在職專班 === 99 === According to Fama (1967, 1976), efficient market hypothesis is no longer fitted to the current stock market, therefore the stock prices can not reflect the real value of respective companies. Nevertheless, investors are able to gain excess returns through various approaches to predict stock prices. In reference to the article of Wyckoff (1970), the financial indices and technical indices are adopted as relative variances. The period of study span is from the year of 2000 to 2009. This study uses the consolidated financial data of underlying stocks of Taiwan 50 ETF and Taiwan 100 ETF as sample collection. Build up a forecasting model through the application of back-propagation network to analyze the effect of financial and technical indices and forecast the stock returns. Utilize predicted hit ratio (PHIT) to determine the validity of selected model and choose the best result for further study of investment simulation. Two investment strategies are developed. Strategy 1 is to buy the stocks which prices are predicted to go upwards. Strategy 2 is to sell the stocks which prices are estimated to decline in addition to buy the stocks which prices are predicted to go upwards. The study also compares with the model of fundamental indices (RAFI) on the validity and investment result.
The empirical study shows:
1. Financial and technical indices have significant effects to stock returns. The R-square of stepwise regression has reached to 86.19%, it’s well matched. On the contrary, RAFI has no significant effect to stock returns. The R-square of stepwise regression only gets 5.37%.
2. The PHIT from the model of back-propagation network has reached to 90.91% and the result from RAFI also hits 67.33%. This reveals the model of artificial neural network has excellent performance to forecast stock returns.
3. The simulation result as follow, the investment combinations organized by strategy 1 and 2 have the average annual returns of 86.19% and 60.08% respectively; Sharp ratios are 1.20 and 0.96 as well. Meanwhile, RAFI gets the average annual return of 44.21% and 18.05% respectively; Sharp ratios are 0.61 and 0.21 separately. On the other hand, the average annual return in the market is 36.96% and sharp ratio is 0.47. In all, the performance from the model of back-propagation network is ahead of the results from RAFI and the market.
|
author2 |
Goo, Yeong-Jia |
author_facet |
Goo, Yeong-Jia Chen, Yen-Chen 陳鄢貞 |
author |
Chen, Yen-Chen 陳鄢貞 |
spellingShingle |
Chen, Yen-Chen 陳鄢貞 Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
author_sort |
Chen, Yen-Chen |
title |
Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
title_short |
Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
title_full |
Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
title_fullStr |
Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
title_full_unstemmed |
Application of Artificial Neural Network Forecasting Model of Financial and Technical Variables on Stock Returns |
title_sort |
application of artificial neural network forecasting model of financial and technical variables on stock returns |
publishDate |
2011 |
url |
http://ndltd.ncl.edu.tw/handle/47176438969286790769 |
work_keys_str_mv |
AT chenyenchen applicationofartificialneuralnetworkforecastingmodeloffinancialandtechnicalvariablesonstockreturns AT chényānzhēn applicationofartificialneuralnetworkforecastingmodeloffinancialandtechnicalvariablesonstockreturns AT chenyenchen yǐcáiwùzhǐbiāojíjìshùzhǐbiāojiàngòugǔjiàyùcèmóxínglèishénjīngwǎnglùmóxíngzhīyīngyòng AT chényānzhēn yǐcáiwùzhǐbiāojíjìshùzhǐbiāojiàngòugǔjiàyùcèmóxínglèishénjīngwǎnglùmóxíngzhīyīngyòng |
_version_ |
1718112115761348608 |