5 ms·
Why use learning when you can fit?
- enriquto 7y agoThis question makes no sense to me. Isn't learning a form of fitting?
- aisofteng 7y agoIn machine learning, a learned model should be able to generalize to unseen data. Curve fitting does not have that characteristic.
- kranner 7y agoCurve fitting does generalise to unseen data, and can even do it well if the training data was representative of the test data.
- enriquto 7y agoI do not understand what you mean. The whole point of curve fitting is to evaluate your curve at points that were not seen.
- 8note 7y agoas long as you're interpolating, it should do just fine. if you're extrapolating, it'll depend on whether you've assumed the right boundary conditions, but that's true for any model
- enriquto 7y agoThe difference between interpolation and extrapolation is not that relevant in dimensions higher than 1.
- darsnack 7y agoYou’re right, but that is the motivation for the blog post. Learning is a form of fitting that performs better in practice than the curve fitting mentioned in the blog post when the required fit is nonlinear and/or high dimensional. The author is making the point that if the model is simple, since learning is just a form of fitting, why not directly apply standard curve fitting techniques?
- joycian 7y agoI don't get this article? Fitting the polynomial is learning, right? (fancy machine learning methods are also not much more than curve fitting)