In many data mining/machine learning books and articles, nobody is mentioning probability distribution of input data and relation between data mining models and distributions.

Does it mean, that it is not important, when applying, for example, a random forest classifier?

When building data mining models, should one consider probability distribution? If yes, how?


By data mining models i mean more like Machine Learning algorithms, as certain data mining models can make assumption about distributioun as Underminer mentioned

  • $\begingroup$ Do you mean probability distribution? $\endgroup$
    – Sycorax
    Apr 15, 2016 at 17:56

1 Answer 1


In general, machine learning classifiers (e.g. random forest) are nonparametric models; that is, they do not make underlying assumptions about the probability distributions of variables being assessed.

Instead, most machine learning classifiers (e.g. random forest, neural nets, SVM, k-nearest neighbor) have algorithmic parameters that are tuned in the training phase, but they don't rely on underlying probability distribution assumptions of the data.

That being said, the term 'data mining models' is very vague, so the above answer is directed at models like you specifically mention in your question, namely the random forest model. There are many parametric statistical models that can be used for classification or other aspects of data mining as well (e.g. logistic regression), but it seems you are referring more towards the algorithmic machine learning approaches.

  • $\begingroup$ Yes, I had in mind more like machine learning approach. I put it in my question. $\endgroup$
    – HonzaB
    Apr 15, 2016 at 18:49

Your Answer

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

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