- 學位論文
利用個股選擇權預測市場下行風險
Using individual stock options to predict S P500 index long and short indicators.
摘要
選擇權自發展以來已成為投資人廣為使用的金融衍生商品之一 ，舉凡債券、 大宗商品、指數到個股，選擇權在商品標的中都扮演重要的角色。過去前人論文指 出個股選擇權價格與成交量資訊與個股未來漲跌表現有統計顯著關係，然而亦指 出指數選擇權價格與成交量與指數未來漲跌並無統計顯著關係。現今科技發展及 其技術也開始應用到金融產業當中，本研究結合資料科學方法以及機器學習模型， 將個股選擇權價格與成交量資訊加總，從下到上組合成一總體市場的選擇權價格 與成交量資訊，並利用 XGBoost、Random Forest 以及 SVC 三種機器學習模型預測 S P500 市場未來漲跌表現，並取得最終預測結果。結果顯示，利用個股選擇權由 下到上組合而成之市場選擇權價格與成交量資訊能夠預測 S P500 指數未來漲跌幅， 當匯集而成之市場選擇權價格以及成交量經過排序後的分數愈高時，S P500 指數 未來跌幅發生機率愈高。尤其發現利用機器學習模型得以提升預測未來漲跌狀況 的精確度，並利用該預測指標控制加減碼最終績效得以擊敗 S P500 指數。
並列摘要
Since its development, option has become one of the financial derivatives widely used by investors. From Bonds to individual stocks, Options plays an important role in financial industry. Previous papers pointed out that the price of individual stock options and trading volume information have a statistically significant relationship with the future performance of individual stocks. However, it is also pointed out that there is no statistically significant relationship between the index option price and index option trading volume have a statistically significant relationship with the future performance of index. Nowadays, technological developments and has also begin to be applied to the financial industry. The research combines data science methods and machine learning models to sum up individual stock option prices and trading volume information, combining the options price and volume information of the overall market from bottom to top, using XGBoost, Random Forest and SVC three machine learning methods to predict the future performance of the S P500 market, and obtain the final prediction results. The results show that the machine learning model can be used to improve the accuracy of predicting index future performance, and the final performance can defeat the S P500 index.