ml-from-scratch

This is a simple machine learning library implementation from scratch only using numpy and standard libraries.

It can be tested by running test files unders tests folder.

Demo

Below two examples of tests. A model tries to learn two different sinusoidal function with one input and one output value. Blue line is the function to be approximated and orange line is model's prediction at each iteration.

sin(x*5)/2+1+x^2;

sin(x*10)/2+1+x^2;