2
$\begingroup$

I would like to know what is the relationship between multivariate analysis and other topics such as: linear regression, neural networks or support vector machines.

According to Wikipedia:

Multivariate statistics is a subdivision of statistics encompassing the simultaneous observation and analysis of more than one outcome variable. The application of multivariate statistics is multivariate analysis.

And for what I know a neural network also allows having multiple outcomes. Also in the same wikipedia page in types of analysis they consider other techiques such as clustering, so what do they have in common multivariate analysis with these data mining techniques? Are one subset of the other or how is it?

I do not know too much about multivariate analysis, but I have learned neural networks, svm and clustering from the Computer Science point of view and not from the statistical one. Any help would be great.

Improve this question
$\endgroup$
1
  • $\begingroup$ It makes more sense to ask "what is the relationship between statistics and machine learning" (which has been asked before on this forum) -- "multivariate analysis" is merely a branch of statistics hence is not at the same level as "machine learning" as a discipline. $\endgroup$
    – Zhanxiong
    Commented Nov 12, 2023 at 15:13

2 Answers 2

Reset to default
1
$\begingroup$

This is somewhat of a philosophical debate without a clear-cut answer.

Some very informed (but not definitive) positions on the matter have been captured in commentary papers such as: Statistical Modeling: The Two Cultures (2001) by Breiman which was likely the first to formally postulate this question, To Explain or to Predict? (2010) by Shmueli or Data science and prediction (2013) by Dhar, that tried to examine the then-emerging Data Science field. At the time of writing this, the single most-upvoted question in CV.SE without an accepted answer is the CV.SE thread: The Two Cultures: statistics vs. machine learning? which is indicative how open-ended and timeless this question is.

As personal anecdote: I learned Principal Component Analysis (PCA) as part of my "Multivariate Statistics" education but nowadays, PCA is almost universally considered as a "core Machine Learning" technique. Does this distinction matter in practice? No.

Improve this answer
$\endgroup$
0
$\begingroup$

You are comparing apples to oranges. Multivariate is when you have more than one predictors. Whereas regression, neural networks are techniques which can handle both univariate and multivariate problems.

Improve this answer
$\endgroup$
3

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.