How can I use a neural network to predict a 1d function?

Hello,

I am running time-costly simulations based on 4 input parameters x=(x1,x2,x3,x4) to calculate a vector with 100 entries y=(y1,...y100). Is there a suitable Matlab function to train a neural network that can predict y(x)? I have had a machine learning class a while ago, so I'm now looking for a good toolbox to start with.

Thanks a lot in advance :)

Answer 1

[James Gross]

Hi Moritz,

We have many different functions and tools for training neural networks on 1D data. However, exactly which function you should use depends on the type of data (e.g. tabular vs timeseries) and complexity of the problem. From the information provided, it looks like you are dealing with tabular data, so I would recommend using the fitrnet or fitcnet functions, as these are generally the most suitable functions for this type of data.

https://www.mathworks.com/help/stats/fitrnet.html

https://www.mathworks.com/help/stats/fitcnet.html

I hope this information helps!

James