透過您的圖書館登入
IP:44.200.145.114
  • 學位論文

統計模型應用於行為股票投資

Applying Statistical Models to Behavioral Stock Investment

指導教授 : 張國華

摘要


以往大部份的人都把所有資金放在銀行裡生成利息,但隨著物價不斷的上漲,人們開始選擇不同的投資方法來賺取額外的收入,而股票是其中最容易入門的投資方式。但在投資股票的過程中,可能因為一些不確定因素,造成投資人的不理性行為,使得投資者進行股票買賣時作出錯誤的決策。而有些人則透過分析這些受不理性行為而影響的股票,從中觀察不理性行為與股票之間的因果關係,將可以找出股票的未來走向。 本研究透過建立操作型定義(Operational Definition, OD)觀察股票產生不理性行為的現象,藉由股票的歷史數據可以找到股票受不理性行為影響的起因,和經過多久時間股價會回升至正常價格,以及相對正面效應的機率,將這些受不理性行為影響的股票稱為行為股票(Behavioral Stock, B-stock),可在前一天將預計明天會正面效應行為股票作為投資標的物。本研究以安全第一(Safety First)為基礎建立投資組合最佳模型,此投資組合最佳模型使用行為股票報酬率的情境,將藉由行為股票反應時間的報酬率與大盤相應時間的報酬率進行迴歸分析,並透過預測大盤報酬率產生符合行為股票隔日正面效應的機率之情境。另外,在選擇標的物時,先以邏輯迴歸(Logistic regression)預測出股票隔日正報酬的機率,透過比較預測出隔日的正報酬機率與操作型定義中股票受影響的機率,篩選出符合預期的行為股票。實驗結果顯示,投資組合明顯優於大盤,顯示本研究的股票投資方法是可行的。

並列摘要


Most people used to put all their money in the banks to generate interests, but as consumer prices kept rising, people started to choose different investment methods to earn extra income, and stocks were the easiest way to get started. However, in the process of investing in stocks, some uncertain factors may cause investors' irrational behavior, which makes investors make the wrong decisions. Previous researches have analyzed the stocks that are affected by irrational behaviors and observe the causal relationship between irrational behaviors and stocks to find out the future direction of stocks. This study observes the phenomenon of irrational behavior in stocks through the establishment of an Operational Definition (OD). Based on the OD and the historical data, we can find the evidences that stocks are affected by irrational behaviors, and estimate how long it takes for stock price back to normal and the corresponding probability of positive effects, we call these stocks affected by irrational behavior stock(B-stock). This research uses stocks that are expected to have positive effects next day as observed at the current day and utilizes the scenario-based safety-first model to obtain the optimal portfolio. In this study, we generate the returns in scenarios using regression analysis on the past data. By predicting the return rate of the market, we generate scenarios that match the probability of the B-stock having positive effect on the next day. In addition, when we select the potential B-stocks, we first use logistic regression to further exam the probability of positive returns next day. By comparing with the nominal probability of positive returns estimated by OD, we filter out those B-stock that meet nominal probabilities. The experimental results show that our investment portfolio is significantly better than the market.

參考文獻


[1] Alexander, G. J., Baptista, A. M., Yan, S. (2017). Portfolio selection with mental accounts and estimation risk. Journal of Empirical Finance, 41, 161-186.
[2] Barber, B. M., Odean, T. (2000). Trading is hazardous to your wealth: The common stock investment performance of individual investors. The journal of Finance, 55(2), 773-806.
[3] Bollerslev, T., Chou, R. Y., Kroner, K. F. (1992). ARCH modeling in finance: A review of the theory and empirical evidence. Journal of econometrics, 52(1-2), 5-59.
[4] Chang, K. H., Young, M. N., Hildawa, M. I., Santos, I. J. R., Pan, C. H. (2015). Portfolio selection problem considering behavioral stocks. In Proceedings of the World Congress on Engineering (Vol. 2, pp. 1-3).
[5] Chang, K. H., Young, M. N., Lin, W. K. (2016). Portfolio Selection Problem Considering Behavioral Stocks under Holding Periods. International Journal of Modeling and Optimization, 6(4), 219-224.

延伸閱讀