Handling missing value
Currently, NumXL allows observations with missing values at either ends of the time series, but not intermediate observations.
In practice, having a sparse runs of observations with missing values is common, so would it be possible if we can use the whole time series, and NumXL will only use the non-missing observations?
By doing so, we avoid any noise produced by the missing values imputation methods.
Comments
Please sign in to leave a comment.