COMP90042, 2014s1: Web Search and Text analysis

Course materials | Questions and answers

Course materials

Week	Lectures	Workshop	Reading
01	Lecture 1a: Administration Lecture 1b: Text, terms, and bags of words Lecture 2: The term-document matrix and the vector-space model	Workshop 1: Terms, vocabularies, and vector-space similarity
02	Lecture 3: Information retrieval Lecture 4: Query expansion	Workshop 2: Information retrieval	Singhal, Buckley, and Mitra, Pivoted Document Length Normalization, SIGIR 1996.
03	Lecture 5: IR Evaluation Lecture 6: Text clustering	Workshop 3: Evaluation and clustering
04	Lecture 7: Singular Value Decomposition	Workshop 4: SVD and LSA
05	Lecture 8: Text Classification Lecture 9: Support Vector Machines	Workshop 5: Text classification
06	Lecture 10: Probabilistic IR Lecture 11: The Two-Poisson Model and BM25	Workshop 6: Probabilistic IR
07	Lecture 12: Language Models for IR Lecture 13: Advanced Uses for Language Models	Workshop 7: Language Models
08	Lecture 14: Probabilistic classification and Naive Bayes Lecture 15: Logistic regression	Workshop 8: Probabilistic classification
09	Lecture 16: Advanced topics in text classification Lecture 17: Probabilistic Topic Models 1: pLSI	Work on second project
10	Lecture 18: Probabilistic Topic Models 2: LDA (part 1)	Work on second project
11	Lecture 19: Probabilistic Topic Models 2: LDA (part 2)	Work on second project
12	Lecture 20: Probabilistic Topic Models 2: LDA (part 3) Lecture 21: Generating Text Lecture 21b: Final Exam	Work on second project

Projects

Project 1

Project 2

Project specification.

Questions and answers

What are expectations around the weekly worksheets?

The primary purpose of the weekly worksheets is to help you familiarize yourself with and extend the material covered in the lectures. The secondary purpose is to give me feedback on whether we're going too fast, too slow, what people are not understanding etc.. Assessment is only a third consideration. If you're finding your having trouble completing the worksheets, or it is taking too much time, then simply document this (in a README.txt file in the Subversion subdirectory), and move on.

How will the weekly worksheets be assessed?

They are assessed on "effort and participation", not outcome. I'm not going to individually mark each worksheet. At the end of the semester, I'll survey the work that each person has done (as located in the Subversion repository), form a conclusion of how much effort they've put in to doing the exercises, and mark accordingly. I'll be providing worksheet answers the week following worksheet submission.

When do worksheets have to be submitted?

Commit to Subversion whatever you've got done by Sunday morning. This will give me a couple of days to see where people are getting up to. Work done after that point can be added when it's done. But it would be defeating the purpose if you let work from one week spill over into the next.

Are the assignments to be done as groups or individuals?

The assignments are to be done as individual work. Only have general conversations with other students about how you might approach them; don't share code, algorithms, written text, etc..

Are the worksheets for groups or individuals?

If you like to work on the worksheets in a group, then I strongly encourage you to do so. Add a note to say who you have worked on your worksheet with. Submit results separately through Subversion. Again, the worksheets are learning, not assessment, resources.

Will lectures be recorded?

Yes, though I'm not familiar with how the system works. Please report to me immediately if you encounter any problems with the recordings.

What if I can't attend a lecture because of a clash with another subject or work?

This does not prevent you from taking the course, but it is your responsibility to read through the lecture slides, listen to the recording, and ask me questions that you have by email.

Last updated: Wed Apr 16 09:05:50 UTC 2014