基于小波变换的LMSV模型变结构研究

提出了应用DWT(离散小波变换)系数累积平方和的LMSV(长记忆随机波动)模型单一变结构点的检测与基于最大重复离散小波变换(MODWT)系数的变结构点的定位方法，并提出了LMSV模型多个变结构点的检测与定位方法．该方法既能对变结构点进行检测和定位，又能同时确定各结构变化发生的尺度．用该方法对沪、深股市综合指数的收益序列进行了波动变结构分析，理论与实证结果表明该方法是有效且可行的．

Study on structure change of LMSV model based on wavelet transformation

The methods for detecting single structure change point using cumulative sums of squares of DWT(discrete wavelet transform) coefficients and locating structure change point using MODWT(maximal overlap discrete wavelet transform) coefficients are proposed. The methods for detecting and locating multiple structure change points of LMSV(long memory stochastic volatility process are also put forward). The methods suggested are proved to be effective and feasible by theoretical analysis as well as the structure change analysis of volatility of the return series of composite indices of Shanghai and Shenzhen stock markets.