Expectation maximization (EM) algorithm implementation using Python.
-
Assume that we have distributions come from two sets of data points, red and blue.
-
From those distribution, we can easily guess the correct estimation of the mean and covariance from each distribution.
- However, if we do not know the correct mean and covariance, we can start from guessing the mean and covariance such as this.
-
Then, we can do the EM algorithm to find the correct numbers. For example, we do in the 10 iterations.
-
After 10 iterations, we can get better numbers for guessing the parameter. We can see it by comparing to the correct numbers.
