我的时间序列被建模为随机游走加回归术语:
#DLM package
model <- function(u) {
local <- dlmModPoly(order = 1, dV = c(1e-06), dW = exp(u[1]))
reg <- dlmModReg(x1, dV = exp(u[2]), addInt = FALSE)
reg2 <- dlmModReg(x2, dV = exp(u[3]), addInt = FALSE)
return(local + reg + reg2)
}
init <- c(rep(0, 3))
outMLE2 <- dlmMLE(data, init, model)
dlmModel <- model(outMLE2$par)
有没有办法获得估计回归系数的t统计量和p值?