How/where to define coefficient for dependent variable in MLR forecast
profile?

Coefficient for dependent variable?

Consider this:

Y = ax1 + bx2 + cx3 + ... (x1,x2 are the independent variables and
Y is the dependant variable.

a,b,c being the coefficients of the independent variables)

This being the standard equation, I dont see what exactly u mean by
the coefiicient of dependant variable.

I have worked on MLR and it follows this equation.

Following is the scenario :

1. Dependent variable Y is forecast key figure ( 9ADFCST)

2. History for Dependent Variable / Key figure Y is 0QUANTITY

3. Dependent Variable X is a key figures 9AADDKF1 and 9AADDKF2

These above parameters can be defined in MLR profile.

So the formula will be

9ADFCST = ( a * 9AADDKF1) + ( b * 9AADDKF2)

Y = ( a * X1
) + ( b * X2 )

Where we have to define the coefficients a and b.

You don't define the coefficients a and b (and the constant c and other
coeffs if you have more independent variables). These are computed
by the MLR engine when you run MLR. Their values can be displayed
by toggling "Para" from the MLR view after running MLR. If you experiment
by changing the historical values of the independent variables and re-running
MLR, you

will see that these coefficients will change (i.e., they are not pre-defined
but rather computed based on the input data).

Note that MLR is not recommended unless you have an expert user who
can interpret what's going on (R**2, Durbin test results, etc.).
Which presents a catch-22 because if you have such users, they would probably
prefer using

other dedicated MLR tools with richer functionality (and ask you to
integrate these to APO!).

Return to :

SAP Hints and Tips on Configuration
and ABAP/4 Programming