0

I am using the KoutParametersEstim() function from the StableEstim package to estimate distributional parameters for the time series RRP shown in the table called SA_PRICE. So when I run the following code:

KoutParametersEstim(SA_PRICE$RRP)$Estim$par

I am receiving this error:

Error in lm.fit(A %*% X, A %*% Y, method = method, ...) : 
  NA/NaN/Inf in 'y'

This is when I apply the function to the full data set. Strangely, the function works well on portions of the data, e.g. the first (or last) 1000 observations. Not sure what the issue is. I do not believe that I have any `NA/NaN/Inf' in my data. It seems something is happening with the computation of the modulus ...

Victor Maxwell
  • 306
  • 1
  • 12
  • Any non-finite values in `A %*% Y` (which is `y` inside `lm.fit`)? – Roland Jul 19 '19 at 06:16
  • Please don't provide data via dropbox (I and many others won't log in there just to answer some question on Stack Overflow). Copy the data into your question body (preferably the output of `dput(yourdata)`). Also, show code that reproduces the error. In short, provide a [minimal reproducible example](https://stackoverflow.com/a/5963610/1412059). – Roland Jul 19 '19 at 06:19
  • I have edited the question to give a better reproducible example. I believe copying the data to the body of the question would get a bit messy because there are almost 9000 observations. I don't believe that you have to login to dropbox to see the file – Victor Maxwell Jul 19 '19 at 07:42
  • Can you add the code to the question, where you read in your data? Where you define SA_PRICE. – grldsndrs Jul 19 '19 at 08:33

0 Answers0