透過您的圖書館登入
IP:3.133.79.70
  • 學位論文

應用遺傳演算法與模糊推論於個股漲跌趨勢之預測

Apply Genetic Algorithm and Fuzzy Inference to Forecast Stock Price Index

指導教授 : 周宗南
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


本研究目的主要探討公司財務指標對公司股價的影響,以台灣經濟新報公布的每日交易財務指標作為研究變數,並透過模糊遺傳演算法、C4.5決策樹及隨機森林,來預測公司財務指標與公司股價漲跌關係。本研究發現,模糊遺傳演算法所預測出整體準確率達64.6%以上,其中以模糊遺傳演算法預測台積電之整體預測準確率高達77.0%,次佳則是隨機森林預測台積電之整體準確率達65.5%;在預測鴻海股價亦為遺傳模糊演算法準確率最高為64.6%。因此歸納出變數進行模糊化和透過基因演變後的預測結果皆比原始變數來的佳,股價的波動性會對預測模型的準確性產生影響。

並列摘要


In the present study, we explore the impact of company’s financial indicators on the companys stock price using the daily data drawn from the database of the Taiwan Economic Journal. Through employing fuzzy genetic algorithm, C4.5 decision tree and random forests, we investigate the relationship between company financial indicators and the companys stock price. We find that fuzzy genetic algorithm predicts an overall accuracy rate of 64.6% or more among them. The prediction accuracy of t on the TSMC’s stock price reached as high as 77.0% by using the fuzzy algorithm. The second best prediction for TSMCs is from the random forest. The accuracy rate is 65.5% overall. As for the prediction of Hon Hais stock price, the fuzzy genetic algorithm could have the highest accuracy rate of 64.6%. Therefore, it seems that the prediction accuracy of model would improve with fuzzifying the variables and the evolution of the genes. Furthermore, the volatility of the stock prices would have the impact on the accuracy of prediction model.

參考文獻


參考文獻
一、 中文部分
1、 王文俊(2017)。認識FUZZY理論與應用(第四版)。新竹:全華圖書。
2、 王惠君(2011)。於模糊理論的時間序列預測模型研究。未出版碩士論文,西北大學,中國。
3、 余尚武、郭至軒(2004)。用基因演算法與模糊理論於台股指數期貨投資策略之研究。未出版碩士論文,國立台灣科技大學資訊管理系。

延伸閱讀