How to use a trained neural network for fitting new data?

I've trained a neural network using a bessel function, and I've gotten a result I'm satisfied with.

Now I want to use the neural network to "smooth the curve" (get rid of the noise)

I am unable to do so, I don't know which function to use, or the correct syntaxis with the variables I have.

Paso 0

clear,clc,close all
load('reto1.mat')

Visualizar los datos

plot(t,S)
hold on
plot(t,I)
hold on
plot(t,R)
hold off

Crear la red neuronal

x = 0:100;
y = real(besselj(0,x));
Red_raza = feedforwardnet([20 10 8 5], 'trainbr');
net.trainParam.show = 1*10^-5;
net.trainParam.lr = 0.5;
net.train.Param.epochs = 1*10^25;
net.train.Param.goal = 1*10;

% net.divideParam.trainRatio = 70/100;
% net.divideParam.valRatio = 15/100;
% net.divideParam.testRatio = 15/100;
net1 = train(Red_raza,x,y);
a = sim(net1,x);
plot(x,a,'o',x,y);

Reto1 is the data I want to use my trained neural network on, which has S, I, and R as dependent variables of t.

I think I can use the neural network to remove the noise on S,I, and R independently and have appropiate curves.

[KSSV]

How Bessel functions and S, I, R related?