【摘要】 基于Kalman滤波理论,考虑财务比率在时间序列上的趋势性和历史数据对结果的影响,构建了财务危机的动态预警模型。首先对动态系统的状态进行描述,建立目标的状态模型,该模型是以时间序列来描述的,此外还建立了财务危机预警的测量方程,利用状态空间法描述目标的状态和测量。然后对Kal-man滤波理论在财务危机预警中的适用性进行分析,利用Kalman滤波器对财务危机预警模型的状态进行Matlab程序计算。并且应用极大似然估计对模型进行参数辨识。采用英国和爱尔兰180家样本公司,5～10年时间序列的财务数据作实证研究,结果表明,由年度破产概率值输出的基于Kalman滤波的动态模型优于静态预测模型的新方法。更多还原

【Abstract】 In this paper we propose a financial distress early warning state space model based on Kalman filter theory according to the trend of time series data and historical data on the results which is creative in the field of finance.First we build the target state space model according to the dynamic system which is described by time series,then we build the measurement equation of financial crisis early warning,and we analyze the applicability of using Kalman filtering to analyze financial distress early warning model.We use Matlab 7.1 to write program and calculate the data.Maximum likelihood estimation is used to parameters identification.The empirical study on 180 British and Irish companies which include 5-10 time series financial data proved that the dynamic model is a more effective algorithm than static models.更多还原