I have a set of input-output data and I would like to derive a mathematical model from that data. I want to know which approach would be better. Should I use NN or should I go for curve fitting? What are the advantages of one over the other?
It depends on the data and wether you need to interpret your mapping function.
If yes, use polynomial curve fitting or similar interpretable methods. Otherwise you can use neural networks or ensemble classifiers like xgboost regressor.
Hello Biswajit,
From your question I assume you have a regression problem. If you have a small low dimensional dataset then you should use curve fitting else NN.
Arseniy, what do you mean by interpreting? Can you please elaborate ? I have already used a polynomial curve fitting once to map my data set but I want to know if NN holds any advantage over it.
Neural network to depend of prestigious computer. And simboliks messages.
And other.
Benedetta Brasca
Ramkrushna, you mean to say that NN should be used for complicated cases involving large data sets with very little correlation ?? Can you please elaborate your answer a bit ? Also can you answer the why?
Neural networks can work better as well, but if you have a small dataset then curve fitting will give good enough results compared to NN, trading off the complexity of the NN and the time it will take to learn.
How can i fit 3d surface for more than one input in NN? ,while plotfit in neural network can only plot single input problems.
- Badges
- Science topic