Does AI Load process understand required order of Regression() formulas?

AI Machine Learning
1

I'm curious how AI Load manages use of LinRegVals(), and the ensuing use of its output variables (slope, R2, SE, etc.). When I use it in an RS I have to run LinRegVals() before using its output variables in a formula. I have several instances of LinRegVals() in my RS, followed by a formula that accesses the output variables. Order is of paramount importance because I'm calculating a regression of several different imported data series, so I run LinRegVals() followed by a formula, then another LinRegVals() followed by a formula.

How does the AI Load process know my required order? That is, how can I be assured that the appropriate LinRegVals() is executed prior to the execution of the associated formula?

I focus on bank stocks. So for example, I run LinRegVals() on the prior 8 quarters of net interest margin, then run a formula against the slope and R2 of the regression output. I then do the same for a bank's Efficiency Ratio. When I create a dataset for AI, I don't get to specify an order-of-execution of the features…..so how does the AI Load process know to run the "NIM" LinRegVals() before my NIM formula, and the Eff. Ratio LinRegVals() before my Eff. Ratio formula?

Thanks.
RJS

2

The typical pattern is to wrap LinReg in Eval;

Eval(LinReg("Close(CTR)", 60), Slope, NA)

That should address any evaluation order concerns.

3

Thanks, Walter. I ran into max. char limits in my formulas so I had to pull the regression formula out of my "regression-assessment" formula. I kinda forgot I'd done that. I'll look at that again and if I can combine my two formulas that will indeed solve my problem.