【作者】 蒲兴成； 汪纪锋； 郑继明；

【Author】 Pu Xingcheng~1, Wang Jifeng~2, Zheng Jiming~1 (1. Computer department, University of Chongqing Post and telecommunication, Chongqing 400065: 2. Communication department, University of Chongqing Post and telecommunication, Chongqing 400065: 3.Automatic department, Chongqing university, Chongqing 400030)

【机构】 重庆邮电学院计算机学院； 重庆邮电学院通信学院；

【摘要】 以改进的相空间重构理论为基础,在假设股票价格变化连续的条件下,提出了一种多元股票价格的预测方法。该方法以不同股票价格在相同时刻的观测值作为相空间的元素,利用嵌入思想,恰当构造相空间的所有元素,再结合改进的相空间重构理论,通过对相空间元素距离的比较,选取最邻近的几个点,根据假设的权值确定方法建立系统的动力系统预测模型,再利用所构建的动力系统模型进行预测,实例结果表明该方法的有效性。更多还原

【Abstract】 On the basis of the theory of the improved reconstructed phase-space, under the proposition that the stock price is continuous, put forward a multi-dimensional predictive method for stock price. The elements of this phase-space is made up of multi-dimensional stock prices in different times, with the embedded ideal, construct all the elements of the phase-space appropriately, combining the improved theory of reconstructed phase-space, through comparing the distance between different elements, select out the most neighbouring elements, construct the impetus predictive model under the given weight decided method, Finally with the constructed impetus predictive model to predict the stock prices. Experimental results demonstrate the effectiveness of the method.更多还原