基于动态交易量预测的VWAP算法交易策略研究

发布时间：2018-03-21 00:54

本文选题：算法交易　切入点：WAP算法　出处：《西北大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着算法交易在金融领域的快速发展和广泛运用,对于交易算法的研究也日益活跃,而作为运用最为广泛的VWAP(交易量加权平均价格)算法交易策略也就成为了研究的热点之一。在VWAP交易策略中,关键的问题在于对交易量分布的预测,预测的准确性直接关系到交易策略的有效性。在历史的VWAP交易策略中,是运用静态的方法,通过历史数据来进行交易量分布的预测,其中并没有将市场的实时信息反映到预测过程中去。本文则通过将股票的实时价格信息引入到交易量分布的预测中去,实现了对交易量分布的动态预测。根据单日内股票交易量通常成“U”型分布的特征,本文首先运用时间序列因素分解的方法,将交易量的历史数据分解成周期因素、趋势因素和波动因素,然后再分别对这些分解出的因素进行预测得到各因素的预测值,进而再将各因素的预测值进行组合得到交易量的预测值,最后再通过引入股票的实时价格信息,对交易量的预测值进行动态的调整,得到最终的交易量预测值。通过数值实验,本文的预测方法比历史的静态预测方法得到的交易量的分布更加接近于市场实际的交易量的分布,证实了本文预测方法的准确性。在交易量分布预测的基础上,本文设计了基于动态预测的VWAP算法交易策略。通过数值实验,本文策略获得的VWAP价格较历史策略获得的VWAP价格更接近于市场实际的VWAP价格；通过在交易相同股票数量的条件下对本文策略获得的VWAP价格与历史策略获得的VWAP价格的差值进行比较,得到了本文交易策略获得的收益高于历史交易策略收益的结果,证实了本文交易策略的有效性。
[Abstract]:With the rapid development and wide application of algorithmic transaction in the field of finance, the research on transaction algorithm is becoming more and more active. As one of the most widely used trading strategies of VWAP (transaction volume weighted average price) algorithm, the key problem in VWAP trading strategy is the prediction of trading volume distribution. The accuracy of prediction is directly related to the validity of trading strategy. In the historical VWAP trading strategy, the static method is used to predict the distribution of transaction volume through historical data. The real-time information of the market is not reflected in the forecasting process, and the dynamic forecasting of the trading volume distribution is realized by introducing the real-time price information of the stock into the forecasting of the trading volume distribution. According to the characteristics of "U" distribution of stock trading volume in a single day, this paper firstly uses the method of time series factor decomposition to decompose the historical data of trading volume into periodic factor, trend factor and fluctuation factor. Then the factors are predicted to get the predicted value of each factor, and then the forecast value of each factor is combined to get the forecast value of trading volume. Finally, the real-time price information of the stock is introduced. The forecast value of trading volume is adjusted dynamically, and the final forecast value of trading volume is obtained. Through numerical experiments, the distribution of trading volume obtained by this method is closer to that of the actual market volume, which proves the accuracy of this method. Based on the prediction of trading volume distribution, this paper designs the VWAP algorithm trading strategy based on dynamic prediction. Through the numerical experiment, the VWAP price obtained by this strategy is closer to the actual VWAP price than the VWAP price obtained by the historical strategy. By comparing the difference between the VWAP price obtained by the strategy and the VWAP price obtained by the historical strategy under the condition of the same number of shares traded, the result that the profit obtained by the trading strategy is higher than that obtained by the historical trading strategy is obtained by comparing the difference between the VWAP price obtained by this paper and the VWAP price obtained by the historical strategy. The validity of the trading strategy is verified.
【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F830.91;F224

【参考文献】