It's a curve fitting problem.
You are given the function generate_observations, that generates n observations
sampled from 2 normal distributions. Each observation generated belongs with probability P
to the first normal distribution.
Write a function estimate_parameters that takes the observations as input and estimates the parameters
of the normal distributions and probability P
Your score will be the total variational distance of your estimated distribution and the original distribution. Get as close as you can to zero.
python optimization.py
pip install -r requirements.txt
Tested on Ubuntu 18.04 LTS, standard laptop equipped with i7 and 16GB of RAM.