Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX.
碩士 === 輔仁大學 === 應用統計學研究所 === 98 === TAIEX has been sold since 1987, and it has been widely accepted by numerous investors. This phenomena has shown that in addition to providing the function of hedging, futures also fit demands such as speculation and price discovery. Since the hedgers and speculato...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2010
|
Online Access: | http://ndltd.ncl.edu.tw/handle/82970270191496474461 |
id |
ndltd-TW-098FJU00506017 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-098FJU005060172015-10-13T18:16:16Z http://ndltd.ncl.edu.tw/handle/82970270191496474461 Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. 類神經網路與支援向量迴歸實作之投資決策模擬—以台灣加權股價指數為例 Li, Yi-Chang 李佾璋 碩士 輔仁大學 應用統計學研究所 98 TAIEX has been sold since 1987, and it has been widely accepted by numerous investors. This phenomena has shown that in addition to providing the function of hedging, futures also fit demands such as speculation and price discovery. Since the hedgers and speculators are the main players in the financial market, investors can obtain the profit of spreading from the target of futures and then advance to buy or sell merchandise for better price. Therefore, some scholars adopted technical analysis for price discovery in literature. Technical analysis is based on stock prices, trading volumes, price and volume changes and other time parameters of the market. These variables act as the basis for investment analysis. In addition to technical analysis, statistical models and machine learning methods are also used for decision making and price prediction. Artificial Neural Network (ANN) and Support Vector Machine (SVM) are the most straightforward and common tools among machine learning methods. This study adopts technical indices and Principle Component Analysis to formulate input variables and put them in the proposed decision support model to help ordinary investors. The performance comparison between the two models built by ANN and SVM respectively and the choice of input variables are the two major topics this study intends to explore. The empirical finding of this study shows that the SVM model with input variables by Principal Component Analysis has the best performance in terms of both prediction accuracy and accumulative profitability. The ANN has better performance when using technical indices as input variables. Dr. Li, Jung-Bin 李鍾斌 博士 2010 學位論文 ; thesis 69 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 輔仁大學 === 應用統計學研究所 === 98 === TAIEX has been sold since 1987, and it has been widely accepted by numerous investors. This phenomena has shown that in addition to providing the function of hedging, futures also fit demands such as speculation and price discovery. Since the hedgers and speculators are the main players in the financial market, investors can obtain the profit of spreading from the target of futures and then advance to buy or sell merchandise for better price. Therefore, some scholars adopted technical analysis for price discovery in literature. Technical analysis is based on stock prices, trading volumes, price and volume changes and other time parameters of the market. These variables act as the basis for investment analysis. In addition to technical analysis, statistical models and machine learning methods are also used for decision making and price prediction. Artificial Neural Network (ANN) and Support Vector Machine (SVM) are the most straightforward and common tools among machine learning methods. This study adopts technical indices and Principle Component Analysis to formulate input variables and put them in the proposed decision support model to help ordinary investors. The performance comparison between the two models built by ANN and SVM respectively and the choice of input variables are the two major topics this study intends to explore. The empirical finding of this study shows that the SVM model with input variables by Principal Component Analysis has the best performance in terms of both prediction accuracy and accumulative profitability. The ANN has better performance when using technical indices as input variables.
|
author2 |
Dr. Li, Jung-Bin |
author_facet |
Dr. Li, Jung-Bin Li, Yi-Chang 李佾璋 |
author |
Li, Yi-Chang 李佾璋 |
spellingShingle |
Li, Yi-Chang 李佾璋 Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
author_sort |
Li, Yi-Chang |
title |
Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
title_short |
Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
title_full |
Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
title_fullStr |
Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
title_full_unstemmed |
Practice of an Investment Decision Support Model by Back-Propagation Neural and Support Vector Regression – an Example of TaiEX. |
title_sort |
practice of an investment decision support model by back-propagation neural and support vector regression – an example of taiex. |
publishDate |
2010 |
url |
http://ndltd.ncl.edu.tw/handle/82970270191496474461 |
work_keys_str_mv |
AT liyichang practiceofaninvestmentdecisionsupportmodelbybackpropagationneuralandsupportvectorregressionanexampleoftaiex AT lǐyìzhāng practiceofaninvestmentdecisionsupportmodelbybackpropagationneuralandsupportvectorregressionanexampleoftaiex AT liyichang lèishénjīngwǎnglùyǔzhīyuánxiàngliànghuíguīshízuòzhītóuzījuécèmónǐyǐtáiwānjiāquángǔjiàzhǐshùwèilì AT lǐyìzhāng lèishénjīngwǎnglùyǔzhīyuánxiàngliànghuíguīshízuòzhītóuzījuécèmónǐyǐtáiwānjiāquángǔjiàzhǐshùwèilì |
_version_ |
1718029239484153856 |