以類神經網路分析財報預測台灣上市公司股價之變動

CCUR > College of Business > Department of Information Management & Graduate Institute of Information Management > Thesis > Item 987654321/23856

Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/23856

Title:	以類神經網路分析財報預測台灣上市公司股價之變動
Authors:	蘇偉庭
Contributors:	資訊管理學系
Keywords:	類神經網路 neural network 財務報表 financial statements 股價變動 stock price fluctuation
Date:	2011
Issue Date:	2012-12-04 10:52:45 (UTC+8)
Abstract:	股票市場中股價的變化是相當快速且難以預測。為了要減少投資失敗的風險，投資人會透過許多分析方法或工具，來預測公司股價在未來的可能變動。許多研究說明財務資訊可以反映公司的真正價值並且反應股價的變動。　　目前財報對於未來股價影響之研究，大多以報酬率的角度來預測未來股價之變動，較少有同時以總體、個股和類股的角度來分析未來短期股價可能變動的研究。故本研究以2005年第2季至2011年第3季台灣上市公司的財報資料與財報公佈前股價表現，分別以倒傳遞類神經網路、廣義迴歸類神經網路和自組織映射神經網路，來建立預測模型，以預測公司發佈財報後未來短期股價變動。並以總體、個股、類股和100年度台灣50的角度來分析預測結果，以提供投資人投資參考。　　實驗結果顯示並非所有台灣上市公司經過預測模型的分析後，都有相當好的預測準確度，較容易受到財報影響股價的公司大多為傳產業的公司，包含水泥、鋼鐵、油電與金融業。投資人可在財報公佈後，關注預測準確度較高的公司，以減少投資失敗的風險。 Predicting stock prices of listed companies in the stock market is not an easy task due to many unpredictable factors could influence stock price fluctuation. 　Many analyzing tools and methods were developed to explain the trend of the stock price based on stock transaction price, volume, financial statements, and other economic information. However, none of the method can predict the trend of stock prices correctly. 　Unlike other researchers using annual return on investment to predict stock price fluctuation, the research applied changing information of incomes, earning per share, net asset value per share in quarterly financial report from 2005 Q2 to 2011 Q3 to analyze stock price fluctuation of individual companies, industries, and Taiwan 50 companies which are larger companies in Taiwan. Three different neural network models are adopted to compare predicting results. 　The experiments showed that not all individual companies have reasonable predicting outcomes but companies of traditional industries such as cement industry, steel and iron industry, oil industry, and financial industry have better results.
Appears in Collections:	[Department of Information Management & Graduate Institute of Information Management] Thesis

Files in This Item:

File	Description	Size	Format
http___thesis.lib.pccu.edu.tw_cgi-bin_cdrfb3_gsweb.pdf		1333Kb	Adobe PDF	945	View/Open

View Licence

Loading...