Suppose we are asked to predict something given a set of features, how do we know if that target is actually predictable? That is, how do we know if there is actually some relation between the dependant and independent features or there are some patterns in the data which could be exploited by a machine learning algorithm?

What if the target outcomes are just random? How do we test for this relationship before we start building ML/DL models?


That would be a part of feature selection. There are many methods to find out if there are relationships between the dependent variable and independent variables. To name a few: plots, measures of correlation, measures of mutual information.

  • $\begingroup$ Thanks . How about cases when you're dealing with unstructured or trace data? For example, if you're trying to predict if someone will perform a certain action, based on their past activities. Let's say, you have a huge volume of trace data, but you don't really know if the next action can be predicted from the current sequence of events. Any idea how we test for relationship in such cases? $\endgroup$
    – Bharathi A
    Jul 18 '21 at 14:11
  • 1
    $\begingroup$ This is a new idea for me. The only thing that does come to my mind (other than empirical methods, of just creating a model) would be to see if an action is more likely to occur after some other action. Something similar to a Markov Chain. $\endgroup$
    – Mateusz
    Jul 18 '21 at 15:05

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.