Given a vector of variables $X=(x_1, x_2, ..., x_n)$ I have previously used genetic programming for deriving an expression for a function $y=f(X)$ from a training dataset of tuples $<X, y>$.

Now, I need a way to learn the target as a (predefined) function of (evolved) functions of inputs. Say $f(X)=\frac{g(X)+b}{h(X)+c}$ is the predefined form of the target function. In this case, The GP evolution searches the space of possible forms of $g(X)$ and $h(X)$ in parallel and finally outputs two best expressions for both of them.

Is any variant of genetic programming proposed in the literature for this scenario?


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.