金融时间序列变点探测的小波模极大值线方法

Outlier's Detection of Financial Time Series Based on Wavelet Modulus Maxima Line Method

研究小波变换方法在金融时序分析中模型变点探测的应用,对金融时间序列采用连续小波变换,通过分析小波变换模极大值线对应的时间序列样本点的小波系数特点,提出了金融时间序列变点探测的小波模极大值线方法,并对广义自回归条件异方差均值模型(GARCH-M模型)进行了仿真计算,其结果验证了此方法的实用性和有效性。该方法更能准确定位金融资产收益率波动所发生的具体时刻,有利于金融资产价格异常时点的正确识别与统计建模分析和资产收益率波动的预测。

This paper investigates the application of wavelet transform methods on outlier detection of financial time series. Through Continuous Wavelet Transform and the analysis of modulus maxima line corresponding sample point, the authors put forward an approach of Outlier＇ s detection of Financial Time Series Based on Wavelet Modulus Maxima Line Algorithms. By digital simulation of GARCH-M model, they prove that the method has much value in practical. The method can be more accurate to identify the concrete outliers of financial assets return what motion take place, and have an important meaning to estimate financial property rate of return as well.