Exercises

Spelling Correction

• Spelling Correction Exercise, Part 1: Error detection and inverted index for n-gram character candidate generation.
• Spelling Correction Exercise, Part 2: Candidate generation (n-gram character and edit based) and scoring with unigram and bigram language models.
• Spelling Correction Exercise, Part 3: TBA, Edit distance and error model computation and use.

Topic Modeling

• Topic Modeling Exercise: Compute topics using Mallet.
• Additional reading/resource: See my post about a Gibbs sampler in R for a toy topic model example, which implements a sampler and runs on the data for the overview paper Probabilistic Topic Models by Steyvers and Griffiths (2007).

Spark

• Spark on AWS, Exercise 1: Launch clusters on Amazon's EC2 service and use Spark to count words in files, and work with HDFS.