Least squares approximation

[Anonymous]

have a poll, and I want to calculate "betas" with the last square aproxximation formula.

I have "Y" (satisfaction), and some other variables as "X"

I have a column with all values of "Y", and one column with all answers of the "X" other answers of the questios of the poll:

Table Name = Registers

Y    X    QuestionCode

3    3           x1

2    3           x2

4    2           x1

      3           x3

      4           x2

"X" and QuestionCode columns are bigger than "Y"column.

I want to acces with a filter on QuestionCode, and multiply Y * (all) x1

That is because of the part of the formula that is SUMMATORY ( X*Y ).

That is differrent of SUMMATORY (X) * SUMMATORY (Y)

What I have been trying is SUMX(Registers[Y] * CALCULATE(Registers, Registers[X]) and filtering on the inform sheet with the QuestionCodes, but when I filter it, it turns "zero"

[Greg_Deckler]

If I understand Least Squares Approximation correctly from here:

http://math4all.in/public_html/linear%20algebra/chapter8.5.html

then you will need to do matrix multiplication. I created a Quick Measure for this:

https://community.powerbi.com/t5/Quick-Measures-Gallery/MMULT/m-p/630231

This uses a technique for Transposing tables in DAX that is documented here:

https://community.powerbi.com/t5/Quick-Measures-Gallery/TRANSPOSE/m-p/630229

Or you could use M to do that job if it does not need to be 100% dynamic.