0

I'm using scikit-learn in Python and I want to use BayesianRidge regression for prediction of a continuous valued target from my continuous inputs. My problem is that I also have a series of binary/categorical inputs and I dont know whether I should still use the BayesianRidge regressor.

If I supply the values as 0 or 1 (or -1, 0, 1) to the BayesianRidge regression, will I get good results? Or is there a better way to do this?

I'm still new to machine learning and I have to admit I find the scikit learn documentation to be overwhelming.

I saw this question regarding a Naive Bayes Classifier, is there a similar approach for Bayesian Ridge Regression?

Mixing categorial and continuous data in Naive Bayes classifier using scikit-learn

Community
  • 1
  • 1
JasonEdinburgh
  • 669
  • 1
  • 10
  • 17
  • 1
    Bayesian ridge regression != naive Bayes. The two are quite unrelated, as naive Bayes isn't even a Bayesian model. – Fred Foo Aug 19 '14 at 13:53
  • Apart from that, this feature representation is in a sense the "preferred" one for linear models, but it depends on the data whether you will get good results. Pick up a textbook such as [*ISLR*](http://www-bcf.usc.edu/~gareth/ISL/) and read up on linear regression. – Fred Foo Aug 19 '14 at 13:58
  • I clearly have very big holes in my understanding. I've seen that book before but was unfamiliar with R so instead made the mistake of buying the "Packt" books on building machine learning systems with Python (in scikit). I also got "artificial Intelligence, a modern approach" by Stuart Russell and Peter Norvig but I had really great difficult with how reliant it was on maths for its explanations in the probabilistic reasoning section. I'm not terrible at maths but I find it a painful way to get an understanding. That book you've recommended looks interesting and may help me with R. Thank You! – JasonEdinburgh Aug 19 '14 at 20:51
  • @larsmans I should ask, Is there a toolkit for python that would allow me to make a bayesian model (or naive bayes) which takes categorical and continous inputs and targets/predicts a continuous output? Similar to how a naive bayes classifier targets/predicts categorical output, but instead with a continuous output. I know I can do this in Matlab with the bayes net toolbox but to be honest, I'm not familiar with matlab and I've just spent months learning python! – JasonEdinburgh Aug 19 '14 at 22:03

0 Answers0