machine learning algorithms predict a mathematical equation with them?

Hi  Yahiaoui,
thanks for reaching out! In a way, all machine learning models we cover here in the program are predicting a mathematical equation. For linear regression, it is a linear equation, where we predict the coefficients. So y = ax + b is the equation, and in a linear regression we predict the slope a and intercept b. We do this based on the data we have.
So, if you have data, which is generated by an equation function, and this can be any equation, not just linear, but say a polynomial (x^3 - 2x^2 + 7) or sinosoid (cos(x)+sin(x)). You can use a machine learning algorithm to determine the equation coefficients. Or it can simply infer the formula and simply predict new results which follow the formula's distribution, without finding it explicitly.
Now, the right choice of a machine learning algorithm will depend on the data you have. If you have a linear problem, then linear regression will be a great choice. If you have more complex data distributions, you can experiment with SVMs and different kernels, to find the right combination which fits the data best.
Hope this helps!

Best,
365 Eli