臺灣博碩士論文加值系統

本文主要應用財經詞彙表，以財報之文字資訊對股價之波動率進行預測。本文利用了一些選取文字特徵之方法，檢視其是否能提升利用財經詞彙表預測波動率之準確度；此外，本文也使用時間序列之GARCH模型，衡量詞彙之特徵是否真能有助於波動率之預測。

The thesis utilize the financial lexicon and text information in Form 10-k to predict the post-event volatility after the release date of Form 10-k. We also propose some methods in feature selection and transformation to imporove the prediction accuracy. Furthurmore, we examine the prediction capability of the selected text features by introducing GARCH.

1 緒論 ..................................1
2 資料集描述 ..................................2
2.1 財報文字資料.................................. 2
2.2 財經基礎詞彙表................................. 2
2.3 波動率計算................................... 4
3 方法 ..................................5
3.1 詞彙擴充 .................................... 5
3.1.1 SimExp................................. 5
3.1.2 SynExp ................................. 6
3.2特徵....................................... 6
3.2.1 特徵選取 ................................ 7
3.2.2 特徵合併 ................................ 8
3.3 預測模型 .................................... 8
3.3.1 支援向量迴歸.............................. 9
3.3.2 GARCH................................. 10
4 實驗................................... 11
4.1 資料前處理................................... 11
4.2 實驗流程 .................................... 11
5 分析結果....................................14
5.1 波動率預測................................... 14
5.1.1 未經擴充之詞彙表v.s.擴充後之詞彙表 ............... 14
5.1.2 特徵合併 ................................ 15
5.1.3 不含詞性資訊之特徵vs.含詞性資訊之特徵 .............................. 16
5.2 對於詞彙預測能力之進一步評估 ....................... 16
6 結論 18
6.1 總結....................................... 18
6.2 未來展望 .................................... 18
Appendices 21
Appendix 1 ....................................... 21
Appendix 2 ....................................... 24
Appendix 3 ....................................... 24

[1] Boyi Xie et al. “Semantic Frames to Predict Stock Price Movement”. In: Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2013, pp. 873–883.
[2] Jasmina Smailović et al. “Predictive sentiment analysis of tweets: A stock market application”. In: Human-computer interaction and knowledge discovery in complex, unstructured, Big Data. Springer, 2013, pp. 77–88.
[3] William Yang Wang and Zhenhao Hua. “A semiparametric gaussian copula regres- sion model for predicting nancial risks from earnings calls”. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2014, pp. 1155–1165.
[4] Ming-Feng Tsai, Chuan-Ju Wang, and Po-Chuan Chien. “Discovering nance key- words via continuous-space language models”. In: ACM Transactions on Manage- ment Information Systems (TMIS) 7.3 (2016), 7:1–7:17.
[5] Tim Loughran and Bill McDonald. “When is a liability not a liability? Textual analysis, dictionaries, and 10-Ks”. In: The Journal of Finance 66.1 (2011), pp. 35– 65.
[6] Tomas Mikolov et al. “Efficient estimation of word representations in vector space”. In: arXiv preprint arXiv:1301.3781 (2013).
[7] Martin F Porter. “An algorithm for suffix stripping”. In: Program 14.3 (1980), pp. 130–137.
[8] Navid Rekabsaz et al. “Volatility Prediction using Financial Disclosures Sentiments with Word Embedding-based IR Models”. In: Proceedings of the 55th Annual Meet- ing of the Association for Computational Linguistics (Volume 1: Long Papers). 2017, pp. 1712–1721.
[9] Understanding Support Vector Machine Regression. The MathWorks, Natick, MA, USA. R2018a.
[10] Ruey S Tsay. Analysis of Financial time series. Vol. 543. John Wiley & Sons, 2005.