EM- Expectation maximization algorithm and applications
EM algorithm is an iterative method for finding maximum likelihood or Maximum A Posteriri (MAP) extimates of parameters in statistical models where the models depends on unobserved latent variable. EM is used to find maximum likelihood estimates given incomplete samples.
EM algorithm applications
- Data clustering in machine learning
- Computer vision
- Natural Language Processing (NLP)
- Probabilistic Context free grammars
- Item response theory models
- Estimating the parameters of Hidden Markov Model (HMM).