Master's course of Prof. Ido Dagan at Bar-Ilan University on the "language" of probabilistic modeling - basic models, estimation and learning methods. Including: probability concepts and definitions, Maximum Likelihood Estimation, hidden variables, Bayesian models, classification, clustering, Expectation-Maximization algorithm and information theory basics.
Second Assignment in 'Applied Probabilistic Models':
In this exercise we explored different methods to estimate the probability of seen and yet unseen events. We used a development set to learn different language models and compare them according to their perplexity on a test set. We also implemented two smoothing methods:
- Lidstone model training
- Held out model training
Third Assignment in 'Applied Probabilistic Models':
In this exercise we implemented the EM clustering algorithm for unsupervised classification of articles into clusters.
Score for both exercises: 100