非平稳序列的随机分析讲义.ppt

library(tseries) mar=arma(Z1,order=c(13,0),include.intercept = T) summary(mar) zr=resid(mar) Box.test (zr, lag = 6, type = Ljung) Box.test (zr, lag = 12, type = Ljung) ct=abs(summary(mar)$coef[,1])/summary(mar)$coef[,2] 1-pnorm(ct) par(mfrow=c(3,1)) ts.plot(zr[-which(is.na(zr))]) acf(zr[-which(is.na(zr))]) pacf(zr[-which(is.na(zr))]) n - 1100 a - c(0.1, 0.5, 0.2) # ARCH(2) coefficients e - rnorm(n) x - double(n) x[1:2] - rnorm(2, sd = sqrt(a[1]/(1.0-a[2]-a[3]))) for(i in 3:n) # Generate ARCH(2) process { x[i] - e[i]*sqrt(a[1]+a[2]*x[i-1]^2+a[3]*x[i-2]^2) } x - ts(x[101:1100]) par(mfrow=c(2,2)) acf(x);pacf(x) pacf(x^2); acf(x^2) x.arch - garch(x, order = c(0,2)) # Fit ARCH(2) summary(x.arch) # Diagnostic tests plot(x.arch) xr=summary(x.arch)$residuals par(mfrow=c(2,2)) acf(xr);pacf(xr) acf(xr^2); pacf(xr^2) data(EuStockMarkets) dax - diff(log(EuStockMarkets))[,DAX] acf(dax);pacf(dax) pacf(dax^2); acf(dax^2) dax.garch - garch(dax) # Fit a GARCH(1,1) to DAX returns summary(dax.garch) # ARCH effects are filtered. However, plot(dax.garch) # conditional normality seems to be violated xr=summary(dax.arch)$residuals par(mfrow=c(2,2)) acf(xr);pacf(xr) acf(xr^2); pacf(xr^2) ARCH模型拟合 定阶：GARCH(1,1) 参数估计：极大似然估计 拟合模型口径：AR(2)-GARCH(1,1) 模型检验 检验方法：正态性检验 假设条件： 检验统计量 检验结果 拒绝原假设 接受原假设 例11续 对美国1963年4月——1971年7月短期国库券的月度收益率序列使用方差齐性变换方法进行分析 假定 函数变换 对数序列时序图 一阶差分后序列图 白噪声检验 延迟阶数 LB统计量 P值 6 3.58 0.7337 12 10.82 0.5441 18 21.71 0.2452 y1-scan() 0.00238 0.00238 0.00236 0.0025 0.00254 0.0026 0.00285 0.00281 0.00241 0.00288 0.00287 0.00292 0.00294 0.00273 0.00271 0.00282 0.00267 0.00273 0.00293 0.00285 0.00296 0.00281 0.00326 0.00321 0.00315 0.00319 0.00313 0.00313 0.00319 0.00313 0.0033 0.00319 0.00315 0.00355 0.0037 0.00371 0.00364 0.00381 0.00372 0.00368 0.00374 0.00389 0.00415 0.00389 0.00343 0.00377 0.00368 0.00364 0.00338 0.00283 0.00271 0.003 0.00309 0.00317 0.00343 0.00347 0.00355 0.0036 0.00398 0.00385 0.00389 0.00444 0.00453 0.00444 0.00432 0.00406 0.00432 0.