Class meeting for 10-605 Rocchio and Hadoop Workflows

This is one of the class meetings on the schedule for the course Machine Learning with Large Datasets 10-605 in Fall_2015.

Slides

Workflows for Hadoop:

• Workflows for Hadoop - Powerpoint, PDF
• The phrases example:
  • PIG source code
  • Java source code
• Some other examples:
  • Naive Similarity Join
  • Optimized Similarity Join

Rocchio:

• Rocchio - Another Fast Streaming Learning Algorithm - PPT, PDF

Also:

• My comments on debugging PIG.

Readings

• Pig: none required. A nice on-line resource for PIG is the on-line version of the O'Reilly Book Programming Pig.

Readings for the Class

• Introduction to Information Retrieval, by Christopher D. Manning, Prabhakar Raghavan & Hinrich Schütz, has a fairly self-contained chapter on the vector space model, including Rocchio's method.

Also discussed

• Joachims, Thorsten, A Probabilistic Analysis of the Rocchio Algorithm with TFIDF for Text Categorization. Proceedings of International Conference on Machine Learning (ICML), 1997.
• Relevance Feedback in Information Retrieval, SMART Retrieval System Experiments in Automatic Document Processing, 1971, Prentice Hall Inc.
• Schapire et al, Boosting and Rocchio applied to text filtering, SIGIR 98.
• Littlestone, Learning quickly when irrelevant attributes abound: A new linear-threshold algorithm, MLJ 1988. Includes the mistake-bound theory.

Things to Remember

• The TFIDF representation for documents.
• The Rocchio algorithm.
• Why Rocchio is easy to parallelize.