0

I would like to estimate the parameter 'r' using GMM, based on a rolling data set [1:2567],[2:2568]... and so on. Finally, the results are to be filled in a matrix. I had already tested the core codes as follows, which yield valid answers.

    x<-Source[1 : 2567,"AAA"]
      z<-as.matrix(x)
      e<-function(r,x){
+         m<-exp(-x/r)-1
+         return(m)
+     }

   coef(gmm(e,z,t0=1,method="BFGS",control=1e-12))
Theta[1] 
1.096466 



     x<-Source[2 : 2568,"AAA"]
      z<-as.matrix(x)
      e<-function(r,x){
+         m<-exp(-x/r)-1
+         return(m)
+     }

   coef(gmm(e,z,t0=1,method="BFGS",control=1e-12))
Theta[1] 
1.102329

However, when I want to do so using a loop structure, it returns with an error message. 

n <- 2
  result <- matrix(rep(0, n), nrow = n)
  for(i in c(1 :n)){
+     x<-Source[i : i + 2566,"AAA"]
+     z<-as.matrix(x)
+     e<-function(r,x){
+         m<-exp(-x/r)-1
+         return(m)
+     }
+     
+     result[i,1] <- coef(gmm(e,z,t0=1,method="BFGS",control=1e-12))
+ }

Error in ar.ols(x, aic = aic, order.max = order.max, na.action = na.action, : 'order.max' must be < 'n.used'

Could anyone help me figure out what had gone wrong? Many thanks~!

(*The data set "Source" is a 5200*4 matrix, with actual data shown in the link below: https://docs.google.com/spreadsheets/d/1AnTErQd2jm9ttKDZa7On3DLzEZUWaz5Km3nKaB7K18o/edit#gid=0)

LINDA HWUNG
  • 11
  • 4

1 Answers1

0

Your error comes from a bad syntax. I added the variable j <- i + 2566. With the following code:

Source <- read.csv(file="source_data.csv", header=TRUE, sep=",") #read csv
n <- 2
  result <- matrix(rep(0, n), nrow = n)
  for(i in c(1:n)){
     j <- i + 2566
     x<-Source[i:j,"AAA"]
     z<-as.matrix(x)
     e<-function(r,x){
         m<-exp(-x/r)-1
         return(m)
     }
     result[i,1] <- coef(gmm(e,z,t0=1,method="BFGS",control=1e-12))
 }
result

I get the following output:

         [,1]
[1,] 1.096466
[2,] 1.102329
Etienne
  • 108
  • 12