[Andrew Ross]

Is there a preferred method for computing extrapolated y values for a linear trendline? Or, for x values other than those in the dataset? I tried doing something like linRegrPredicted(Year,MyYvalue) and then adding more rows to the dataset with the new x values I wanted, but nothing for MyYvalue. It didn’t report any results for linRegrPredicted, though.

I did manage to use something like this:

linRegrSlope(year,MyYvalue)*year+linRegrIntercept(year,MyYvalue)

and added rows to the dataset at the x values I wanted, but that feels a little inelegant. It can also give the impression that those x values were in the original dataset, but just happened to not have y values.

Is there a preferred way to do this?

[Bill Finzer]

Hi Andrew,

Thanks for this post. The “preferred” way would be as you describe in your first attempt. But, as you note, that doesn’t work. So I’m going to add that as a bug report this morning and hopefully get it fixed and released by end of next week.

I don’t have a good suggestion for generating predictions for values not in the dataset. Will have to think about. Any ideas?

Bill

[Author]

Posts