影響台灣金融類股股票報酬因素之研究

A Study on the Determinants of the Financial Stock Returns in Taiwan

鄭淑燕(SHU-YEN CHENG)

指導教授 ： 陳建宏

朝陽科技大學/管理學院/財務金融系/碩士(2019年)

摘要

物價扶搖直上，薪資停滯國民所得下滑，增加家庭收入勢在必行，投資理財成為全民運動。若能運用財務變數，準確預測金融類股股票報酬率，對投資人來說：可降低風險提升獲利。本研究以台灣經濟新報資料庫為基礎，進行資料採礦，挑選10個條件屬性和1個決策屬性-股票報酬率。主要實證結果如下:(1)應用屬性選取，找出重要條件屬性為：每股盈餘、稅後淨利率、營收成長率、淨值成長率。(2)衡量各預測模型之準確率，不論有無屬性選取，皆以決策樹J48的準確率最高。(3)Naïve Bayes、IBK、J48 三種不同分類器，有屬性選取的模型準確率皆優於無屬性選取的模型。

關鍵字

決策樹 ； 財務變數 ； 資料採礦 ； 股票報酬率

並列摘要

This study conducts data mining to find the determinants of the financial stock returns in Taiwan. We select 10 financial ratios as conditional attributes and select stock return rate as decision attribute from the Taiwan Economic Journal. The main empirical results are as follows: (1) Apply attribute selection to find out the important conditional attributes are earnings per share, net profit after tax, revenue growth rate, and net asset value growth rate. (2) Measure the accuracy of each prediction model, with or without attribute selection, the decision tree J48 has the highest accuracy. (3)Among the three different classifiers: Naïve Bayes, IBK, and J48, the accuracy of the model with attribute selection is better than the model without attribute selection.

並列關鍵字

Decision Tree ； Financial Variables ； Data Mining ； Stock Returns

參考文獻

英文文獻

1. Altman, N. S. (1992). An introduction to kernel and nearest-neighbor nonparametric regression. The American Statistician, 46(3), 175-185.

2. Berry, M. J., & Linoff, G. (1997). Data mining techniques: for marketing, sales, and customer support. John Wiley & Sons, Inc.

3. Chen, Y. S., Tsao, H. H., & Huang, C. H. (2017). Performance Evaluation of Comprehensive Function Classifiers: Taking Growth Rate Data as an Example. Applied Science and Management Research, 4(1), 61-70.

4. Chen, L. S., & Liu, C. C. (2015). Using feature selection approaches to identify crucial factors of mobile advertisements. In Proceedings of the International MultiConference of Engineers and Computer Scientists (Vol. 1).