16. EM (Expectation Maximization) algorithm EM algorithm is an iterative optimization method that finds the maximum likelihood estimate (MLE) of parameters in problems where hidden/missing/latent variables are present. Suppose you have a data set with n number of data points. Clustering is the task of finding out k natural groups for your data when you dont know (or dont specify) the real grouping. Such clustering problem can be tackled by several types of algorithms, e.g., combinatorial type such as k-means or hierarchical type such as Wards hierarchical clustering. However, if you believe that your data could be better modeled as a mixture of normal distributions, you would go for Gaussian mixture model (GMM). Programmatically implement Unsupervised learning using EM+GMM.

Demand:

1. Describe corresponding strategy.

2. Output the correct answer.

3. Analyze the time complexity of the algorithm.