JP_LM Model

# Cut-and-paste code below into window above and Run

#

# Japan (JP) Model (1950-2000)

#

#

 #Measurement Matrix   (Growth) (XREAL-HOURS) (HOURS + XREAL - N - U)

 #

#           Q      N      U  HOURS XREAL      L

#[1,]  0.4120  0.411  0.412  0.402 0.400  0.412

#[2,]  0.1811 -0.194  0.122 -0.638 0.694 -0.161

#[3,] -0.0347 -0.563 -0.290  0.620 0.439 -0.143

#

# Fraction of Variance 

#

 #Fraction of Variance 

#[1] 0.995 1.000 1.000

#

merge.forecast <- function (fx,n=1) {

x <- splice(fx$pred,fx$forecast[[n]])

colnames(x) <- seriesNames(fx$data$output)

return(x)

}

AIC <- function(model) {informationTestsCalculations(model)[3]}

require(dse)

require(matlab)

f <- matrix( c(   0.9946733059, -0.06609345, -0.06755142,  0.15827394,

                         0.0224205739,  1.06706048,  0.04207700,  0.01267197,

                        -0.0002452678, -0.06837779,  0.99535499, -0.01042065,

0.00000000,  0.0000000,  0.0000000,  1.0000000000

),byrow=TRUE,nrow=4,ncol=4)

#

# To stabilize model, uncomment next line

# f[1,1] <- f[2,2] <- f[3,3] <-  .9

#

#

h <- eye(3,4)

k <- (f[,1:3,drop=FALSE])

JP_LM <- SS(F=f,H=h,K=k,z0=c(0.15827394,  0.01267197 , -0.01042065,  1.0000000000),

              output.names=c("JP1","JP2","JP3"))

print(JP_LM)

is.SS(JP_LM)

stability(JP_LM)

JP_LM.data <- simulate(JP_LM,sampleT=100,start=1950)

#JP_LM.data <- simulate(JP_LM,sampleT=100,noise=matrix(0,100,3),start=1950)

JP_LM.f <- forecast(l(JP_LM,JP_LM.data),horizon=150)

JP_LM.fx <- merge.forecast(JP_LM.f)

tfplot(JP_LM.f)

AIC(m <-  l(JP_LM,JP_LM.data))

JP_LMx <- SS(F=f,H=h,Q=eye(4,3),R=eye(3,3),z0=c(0.15827394,  0.01267197 , -0.01042065,  1.0000000000),

              output.names=c("JP1","JP2","JP3"))

     shockDecomposition(JP_LMx)