Function Curve Classification

I have a 1D function changing shape (linear to S-shaped or vice-versa) depending on the input. I want to perform some training on a set of samples with ground-truth data and be able to classify the unknown input of the function in two distinct classes.

I am looking for possible ways (including current state-of-the art) to approach this problem. Any suggestion?

In particular I am wondering if I should consider the set of discrete points generated by the functions as vector features and simply put everything into SVM or similar or given the fact those samples are an ordered sequence other methods might be more suitable.

asked Oct 18 '13 at 20:40

memecs
1●2●2●3

edited Oct 18 '13 at 20:42

Can you be more explicit? What kind of input do you have and what kind of output do you have? An SVM will take an F dimensional features and give you a label. Is your input to the SVM the function shape and the output the unknown input?

(Oct 18 '13 at 21:41) Leon Palafox ♦

Subscription:

Once you sign in you will be able to subscribe for any updates here

Tags:

classification ×237
graph ×7
function ×6
curve ×5

Asked: Oct 18 '13 at 20:40

Seen: 379 times

Last updated: Oct 18 '13 at 21:41

Function Curve Classification

Subscription:

Related questions