0
$\begingroup$

Let's suppose that I have some data, and I have a vector representation of each data point. For example, one data point might look like this: [0, 1, 0, 3, -2, 2.3]. Now suppose that for each vector, the value in dimensions 1, 2, and 3 can only ever be 0 or 1. Then, the value in dimension 4 can be any positive integer value. Then the value in the 5th dimension can be any integer, and the value in the 6th dimension is a decimal value.

Can machine learning be used on data of this form? Or does every dimension need to be the same kind of number, and have the same range (such as all dimensions are binary, or all dimensions are integers.)

$\endgroup$
1
$\begingroup$

You can certainly use machine learning (and statistics) on such data! For better advice ask about your real problem, but you will find many examples on this site with mixed data, you could start by looking through the tag .

$\endgroup$
1
  • 1
    $\begingroup$ One example of a mixed-data regression is an ANCOVA model. (I would expect to use that for inference rather than prediction, but it could be a simple machine learning predictive model.) $\endgroup$ – Dave Nov 23 '20 at 17:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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