基於創新的風險評估策略並結合演化計算解決投資組合最佳化問題

Portfolio Optimization Based on Novel Risk Assessment Strategy with Evolutionary Algorithm

廖柏宇(Bo-Yu Liao)

指導教授 ： 周耀新

國立暨南國際大學/科技學院/資訊工程學系/碩士(2015年)

https://doi.org/10.6837/NCNU.2015.00231

摘要

在股票市場中投資，首要面對的就是選股問題，而如何挑選出一個兼顧低風險且高報酬投資組合是個值得探討的問題。在計算投資組合風險上，傳統的方法計算複雜度高，且無法計算變異係數，以至於無法準確評估投資組合風險，且無法完全代表投資者對風險的看法。
本研究提出一種全新計算投資組合風險的算法，我們利用整個投資組合資金水位的變動來表示其風險，不僅計算容易，更能準確評估整體投資組合的風險且資金水位的變動非常能表示投資者心情的浮動。並利用演化計算結合夏普指標也就是每單位風險分之報酬去搜尋-低風險且穩定報酬的投資組合。
資金分配方面，本研究針對平均分配以及比例分配分別做測試，並觀察投資組合組成以及表現。
此外本研究用滑動視窗的方法去避免股票領域常見的過度適應的問題，以及測試各種訓練、測試週期對投資組合的影響。
從實驗結果可以得知，在與傳統的計算風險的方法比較，我們的方法能夠找到最佳的投資組合。

關鍵字

投資組合 ； 變異係數 ； 夏普值 ； 資金水位 ； 演化計算 ； 滑動視窗

並列摘要

Stock selection is an important issue when it comes to investing in the stock market. However, it is worth investigating the problem of selecting portfolios while considering not only low risk but also high return on investment. The calculation process of the traditional method is highly complex and Coefficient of Variation (CV) cannot be calculated. So the process of the traditional method unable to accurately assessment of portfolio risk. Hence, this study proposes a new method to calculate portfolio risk. We utilize funds standardization in order to consider the risk of a portfolio and drastically reduce computation complexity and accurate assessment of portfolio risk. Funds standardization is able to represent fluctuations of investor mood. Moreover, using an Evolutionary Algorithm combined with the Sharpe Ratio is able to identify the low risk and stable returns of a portfolio. Distribution of Resources, the study were test for equal distribution and allocation of funds, and to observe the portfolio composition and performance. Moreover, over-fitting is a common problem in the stock market, and so this paper uses sliding windows to avoid the over-fitting problem, and tests all kinds of training periods and testing periods that impact on the portfolio. The experimental results show that the proposed method, compared with the traditional method of calculating risk, is able to identify the optimal portfolio and performs efficiently and outstandingly when it comes to this problem.

並列關鍵字

Portfolio ； Coefficient of Variation (CV) ； Sharpe Ratio ； Sliding windows Funds Standardization ； Evolutionary algorithm

參考文獻

[1] H. Markowitz, ”PORTFOLIO SELECTION,” The Journal of Finance, Vol.7, Issue 1, pp.77-91, Mar. 1952

[2] W. Sharpe, “Investors and markets: portfolio choices, asset prices, and investment advice,” Princeton University Press, 2011.

[3] W. K. Wong, E. Bai and A. W. Chu, ”Adaptive Time-Variant Mod for Fuzzy-Time-Series Forecasting,” Systems, Man, and Cybernetics (SMC), IEEE Transactions on, Vol.40, no.6, pp.1531-1542, Dec. 2010

[4] Y. Leu and T. I. Chiu, "An effective stock portfolio trading strategy using genetic algorithms and weighted fuzzy time series," Nano Information Technology and Reliability (NASNIT), IEEE Transactions on, pp.70-75, Oct. 2011

[5] F. A. de Oliveira, L. E. Zarate, M. de Azevedo Reis and C. N. Nobre, ”The use of artificial neural networks in the analysis and prediction of stock prices,” Systems, Man, and Cybernetics (SMC), IEEE International Conference on, pp.2151-2155, Oct. 2011