I'm new to machine learning so am quite confused with the above concepts. It seems to me both flexibility and complexity measures how well the model fit the data (in terms of the curvy-ness), so what's the difference between the two? In addition, I saw some websites say complexity can be indicated by the number of features in the model; but does 'feature' in this context mean variable, e.g. x and y (two variables), or term, e.g. x and x^2 (two terms), and is it different from 'predictor'? Also, I know flexibility can affect the generalization error, but except that, does flexibility affect any other aspects of model performance? Thanks for answering.

  • 1
    $\begingroup$ Can you include the sources that you refer to? $\endgroup$ – Sammy Jan 24 at 12:31
  • $\begingroup$ The website about complexity is this one: innoarchitech.com/blog/…. $\endgroup$ – capcapuccino Jan 24 at 13:57
  • $\begingroup$ the 3 terms are not identical nevertheless they are related. For instance number fo predictors/features affects complexity, but there are complex functions to be found with only a few inputs. Flexibility relates to how easily the model can generalise and a very complex model may not be flexible, but again there is the other side of this $\endgroup$ – Nikos M. Jan 24 at 16:56
  • $\begingroup$ @NikosM. Thanks for replying. But could you maybe provide some examples for better illustration? I'm not totally sure that I understand the differences here. $\endgroup$ – capcapuccino Jan 25 at 5:22

Your Answer

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

Browse other questions tagged or ask your own question.