Dr. Max L. Wilson

Lecture Schedule

Assignments

Assignment 1 (15%): Due Thu 3rd Nov 2011 at 11am

Building the basics of a search engine

Implement a small search engine application based on TFIDF with the log rule. Download these text files. You should design this as a webpage that can be tested. You can use any language you like. If building a webpage is really a problem, you can produce a seperate application, as long as it can be sent to me. Please contact me if it's a problem.

Queries for the test:

• environmental
• Amazon
• Japan whaling
• Soviet
• business

Results: should look like:

x Matching documents:
Filename.txt: y hits (z.nnnn)
Filename.txt: y1 hits, y2 hits, (z.nnnn)

Where x is the number of documents, y is the number of times a word is in the document. If the search has multiple words, then y1 is the first word and y2 is the second word. Finally, z.nnnnn is a floating point number of the documents TFIDF+logrule score) for the document. Results should be in ranked order, highest first. Do not return results with a score of 0.

Notes:

1. You can easily figure out the answers manually to check your code is working
2. Strip out all punctuation
3. Be case insensitive
4. Here's how to use your uni webspace

To Submit on blackboard: A zipped file that contains 1) an A4 report including a) an explanation of your approach, b) any partiuclar assumptions about your code, such as how you have handled punctuation, c) the results for each query, and d) instructions for running your code (this may be a link to a website, or instructions to compile/run an application). And 2) A folder containing your well-commented code.

Marking: You will be marked on a) accurate ranking scores, b) correct ranking, and c) quality of coding. Part marks will be given for near misses.

Assignment 2 (15%): Due 13th Dec at 11am

How do search services match up to Information Seeking Theory?

Sii (the Search interface inspector) is a new method that provides you with a process for evaluating search interfaces in accordance to some the theories you have been taught in the last few classes. Your task is to evaluate and compare two online fashion retailers (ASOS and very.co.uk) using the Sii method.

Instructions:

1. Register at the Sii website
2. Input the details for the two shopping services.
3. Analyse them both
4. Write a report on the two shopping services, discussing how they compare and how they could be improved.

To hand in: a report, usings screenshots, describing:

1. which features of the websites were included, and which ones were not (and why)
2. 10 examples, described step-by-step, of the moves required to perform a tactic with one feature of one of the services (1 tactic from 5 features of each website NOT 10 tactics from 1 feature of one website)
3. a discussion, based on the graphs, of the strengths and weaknesses of each interface
4. recommendations for improving each interface
5. a description and visualised (through data entry of a third concept design) of how the support would change based on these recommendations
6. a 'meta-method' section with any suggestions for making the Sii method easier to use for the first time

Submission: is via blackboard. Ask in advance if you are not on this system.

Additional notes: Focus only on the fashion shopping i.e. features that help you buy clothes to have them delivered home. Ignore additional searching, like help forums etc. You can have a look at the RYA examples on the Sii website for how the old and new versions of the RYA website were analysed. There are more examples there.

This work is solo. Do discuss, but don't share your analysis links with each other.