Can anyone guide on how to find the relationship between two matrices for prediction? I have 50 psychological variables and 100 movie features. The value of each variable is in the scale of 1-100. If I know the value of the psychological variables I want to predict the value of the movie feature. How do I do that? Does a least square multivariate model give good estimation in my case?
Hi Hetal,
What you need is PLS regression (Partial Least Squares Regression), it allows you to reduce the dimension of a Y matrix and an X matrix at the same time you correlate the newly obtained (so-called "latent") Y and X matrices.
Google it and see if that's what you were looking for. In that case, tell me and I can provide you with some papers that are a good introduction to this method.
Hope this helps.
Kind regards.
Thank you Emili. i will be grateful to you if you could mail me some papers.
Hetal -
Hopefully you have lots of data (N observations, n_training + n_test
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics