standard error of values or some estimator ?

for example  standard error of sample average, if you are estimating your population average using the sample average

In regression, to calculate standard error, you must have historical or true values and predicted values t

You will get the difference between each pair of historical and predicted value

You must square this difference

You must get the sum of the squared differences

You will divide this sum by n-1

You will get the square root of this quotient

You have the standard error

Standard error is not an accepted term in the vocabulary of metrology, because there is no generally accepted definition.  If your measurements are in error, they should be corrected rather than standardized.