管理學院: 財務金融學研究所指導教授: 姜堯民劉冠吟Liu, Kuan-YinKuan-YinLiu2017-03-032018-07-092017-03-032018-07-092016http://ntur.lib.ntu.edu.tw//handle/246246/274148關於財金相關文件或新聞的文字探勘近幾年才逐步開始發展，本論文使用Loughran和McDonald所建立的字典，將美國各公司申請首次公開發行時所提交之公開說明書內的文字依照其語調做分類，探討不同分類的字彙是否會對上市後的股價、公司的營運表現，以及從提交公開說明書至正式上市所花費的時間有所影響。我們發現以公開說明書內的字彙來預測上市後的股價結果可能會不一致；另外，字彙與上市後公司的營運表現，以及從提交公開說明書至正式上市所花費的時間之間並無顯著性的相關。While there exists a substantial amount of literature focusing on how concrete information in IPO prospectus influences IPO pricing, research concerning the relationship between abstract information and IPO pricing is still less and has not yet been developing since recently. I use the financial sentiment dictionary established by Loughran and McDonald to measure tones in the summary section of Form S-1 and examine how documents’ tone influence underpricing, post-IPO operating performance, and the process time between issuing and filing. I find that use word classification from Loughran and McDonald’s dictionary to predict IPO underpricing might have inconsistency and the connection between word in prospectus and first-day return might be nuanced. I also find little statistically significant correlation between tones in IPO prospectus and post-IPO operating performances, and the process time between filing and issuing.論文使用權限: 不同意授權首次公開發行公開說明書文字探勘IPO折價發行IPOstextual analysissentiment analysisunderpricingfilingsthesis10.6342/NTU201600419