双指标极小投资模型的设计与优化

发布时间：2018-05-09 11:38

本文选题：量价关系 + 协整分析　；参考：《华南理工大学》2012年硕士论文

【摘要】：证券技术分析是广泛采用的一种证券投资分析方法。投资者根据技术指标预测股市未来的行情走势以期获得更多的收益，此时，最关键的就是寻找股价反转点。如果能够较早的判断股价的反转点，做到低吸高抛，投资者将获得丰厚的回报。交易量和股价（简称量价）间的关系在技术分析中占据着重要的地位。基于量价，本文设计了一种识别反转点的模型，对于投资者进行更好的投资决策具有参考意义。根据证券技术分析的分析理念，及收盘价和交易量间存在的协整关系，本文将收盘价和交易量同时进行考虑，设计了双指标极小投资模型（简称双极模型）。双极模型中有两个参数，分别是P和Q，其中，P代表当日收盘价是连续P天的极小值，Q代表当日交易量是连续Q天的极小值，不同的参数组合代表不同的投资模型。双极模型的设计思路是给定P和Q，若当日收盘价是连续P天的极小值并且当日交易量是连续Q天的极小值，则双极指标值取为1，否则取为0，当双极指标值由1变为0时，则买进股票；买进股票后，当股价从一个新高位置下跌，从这个新高位置开始，若4天内回落达到5%，则卖出股票。对于股票收盘价序列和交易量序列一阶单整，且差分后序列一阶自相关的情况，理论上证明了双极模型设计的合理性。假定投资期限为n年，利用双极模型进行投资模拟。首先，计算不同参数组合下每次买卖的投资收益，按复利计算其n年投资总收益，并计算其年均收益率。其次，根据EGARCH模型，算出每次投资的日VaR值，从而得到一组VaR序列，将其99%分位数作为投资期限内的总风险值。用年均收益率和总风险值综合评估每次投资的优劣。再次，为了寻找较好的参数组合，分别作出不同参数组合下的收益率等高线和风险等高线，并观察收益率和风险对参数的敏感程度。最后，对参数进行优化，得到最佳参数组合。本文用万科A股票进行实验，结果表明，收益率对P比较敏感，对Q不太敏感，而风险对P和Q都比较敏感；若采用RAROC指标，最佳参数组合为P=2、Q=5；对于不同类型的投资者，根据有效边界和无差异曲线，得到其相应的最佳参数组合，其中，对于保守型投资者来说，最优的参数组合为P=5、Q=4；对于进取型投资者来说，，最优的参数组合为P=3、Q=4。
[Abstract]:Securities technical analysis is a widely used method of securities investment analysis. When investors use technical indicators to predict how stocks will fare in the future in the hope of earning more, the key is to find a price reversal. If you can judge the reverse point of the stock price earlier and sell it low, investors will get a good return. The relationship between trading volume and stock price plays an important role in technical analysis. Based on the quantity price, this paper designs a model to identify the inversion point, which has reference significance for investors to make better investment decision. According to the analysis idea of security technical analysis and the cointegration relationship between closing price and trading volume, this paper considers the closing price and trading volume at the same time, and designs a double-index minimal investment model (dipole model for short). There are two parameters in the bipolar model, one is P and the other is Q.With P representing the closing price of the day is the minimum value of continuous P day, Q represents the minimum value of continuous Q day trading volume, and different parameter combinations represent different investment models. The design idea of the bipolar model is to give P and Q, if the closing price of the day is the minimum value of the continuous P day and the trading volume is the minimum value of the continuous Q day, then the bipolar index value is taken as 1, or 0, when the bipolar index value changes from 1 to 0. You buy stocks; after you buy stocks, you go down from a high, start at that high, and sell if you fall back to 5 in four days. It is proved theoretically that the design of the bipolar model is reasonable for the stock closing price sequence and the trading volume sequence in the case of the first order autocorrelation of the first order of the stock closing price sequence and the trading volume sequence and the first order autocorrelation of the difference sequence. Assuming that the investment period is n years, the investment simulation is carried out by using the bipolar model. First of all, the investment income of each purchase and sale under different parameter combinations is calculated, the total investment income per year is calculated by compound interest, and the average annual return rate is calculated. Secondly, according to the EGARCH model, the daily VaR value of each investment is calculated, and a set of VaR sequence is obtained. The 99quartile is taken as the total risk value within the investment period. The average annual rate of return and the total risk value are used to evaluate the merits and demerits of each investment. Thirdly, in order to find a better combination of parameters, the yield contour and the risk contour under different parameter combinations are made, and the sensitivity of the yield and risk to the parameters is observed. Finally, the optimal combination of parameters is obtained by optimizing the parameters. In this paper, Vanke A stock is used to experiment. The results show that the return rate is sensitive to P, less sensitive to Q, and the risk is sensitive to both P and Q; if the RAROC index is used, the best parameter combination is Pf2Q5; for different types of investors, According to the efficient boundary and the non-difference curve, the corresponding optimal parameter combination is obtained, in which, for conservative investors, the optimal parameter combination is P5 / QF4, and for enterprising investors, the optimal parameter combination is Pf3 / QF4.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.91

【参考文献】