水平窗口能量计算的股市趋势预测算法

发布时间：2018-04-09 14:21

本文选题：能量窗口　切入点：K线特征　出处：《计算机工程与应用》2017年21期

【摘要】：水平趋势持续时间短,方向变化的不确定性大,水平状态下趋势预测成为股市趋势预测的难点。基于水平窗口的能量计算,提出一种水平窗口趋势预测的BP神经网络算法(WE-BPNN)。算法首先给出短线趋势划分标准,在此基础上引入水平窗口定义;然后,通过对K线组合能量和均线组合能量进行量化计算,融合这两种能量得到窗口能量;最后,将窗口能量引入到BP神经网络预测窗口方向。由于能量对于趋势的作用具有滞后性,存在能量蓄而不发的情况,会影响到趋势判断的准确性,因而在WE-BPNN算法的基础上给出引入能量调节因子的BP神经网络(EF-BPNN)算法,动态调整窗口能量因子对于趋势预测的影响权重。在上证数据上的实验结果表明,EF-BPNN算法具有更好的性能。
[Abstract]:The trend of the horizontal trend has a short duration and the uncertainty of the direction change is great. The forecast of the trend in the horizontal state becomes the difficulty of the forecast of the trend of the stock market.Based on the energy calculation of horizontal window, a BP neural network algorithm for predicting the trend of horizontal window is proposed.On the basis of this, the definition of horizontal window is introduced. Then, by quantifying the combination energy of K-line and mean-line combination, the window energy is obtained by combining these two kinds of energy.The window energy is introduced into the BP neural network to predict the window direction.Because of the lag of energy to the trend and the fact that the energy is stored but not generated, it will affect the accuracy of the trend judgment. Therefore, based on the WE-BPNN algorithm, the BP neural network (BP neural network) algorithm which introduces the energy regulation factor (EF-BPNN) is proposed.Dynamically adjust the influence weight of window energy factor on trend prediction.The experimental results on Shanghai Stock Exchange show that the EF-BPNN algorithm has better performance.
【作者单位】：合肥工业大学计算机与信息学院;
【基金】：国家自然科学基金(No.61175051,No.61175033) 国家重点基础研究发展计划(973计划)(No.2013CB329604)
【分类号】：F832.51;TP183

【相似文献】