I just put a bounty on Calculating Linear Discriminant Analysis Coefficients before seeing the ML discussions on Meta, so I apologize for not asking about this beforehand. I'll be sure to make a future note of what comes out of this discussion.

The question, in short, is asking about how to manually calculate the coefficients outputted by a machine learning algorithm using matrices. In addition, R code is a substantial part of this, and it looks like the algorithm itself involves a lot of advanced linear algebra which I am not familiar with (singular value decomposition). Therefore, I don't consider this question to be in the category described in As machine learning gets more programming-oriented, should we get more programming oriented?.

Any thoughts on this?

  • 2
    $\begingroup$ IMHO, it is perfectly on topic. $\endgroup$
    – Tim
    Oct 23, 2017 at 16:12
  • $\begingroup$ I agree with @Tim, your Q is perfectly on topic. $\endgroup$
    – amoeba
    Oct 23, 2017 at 20:46

1 Answer 1


Just containing code doesn't make it off topic; the question seems to me to be firmly on-topic.

In any case, in relation to programming questions (which I don't think this one actually is), the help/on-topic says:

if it needs statistical expertise to understand or answer, ask it here

So even if it were more of a programming-related question than it looks at first read, I think this phrase would mean it would be on topic.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .