Homework 2

Due Date: February 7, 2002

• Geocoder:
  • URL: http://cronus.isi.edu:8080/geocode.html
  • Type: form
  • Input parameters: street address, city, state, zip code
    Example input:
    • street address: 4676 Admiralty Way
    • city: Marina del Rey
    • state: CA
    • zip: 90292
  • Output parameters (to be extracted from answer page):
    For five addresses of your choice, you should extract the following information:
    • Latitute
    • Longitute
    • Error Info
  • Attention: please be patient with the geo-coder!
    (On its best day, it takes about 10 seconds to retrieve a location.)

• CIA World Factbook:
  • URL: http://www.cia.gov/cia/publications/factbook
  • Type: frame-based navigation
  • Input parameters: none
  • Output parameters (to be extracted from site):
    For five countries of your choice, you should extract the following information:
    • Country Name
    • Location
    • Geographic Coordinates
    • Map Reference
    • Area - comparative
    • Climate
    • Terrain
  • Note: in solving this problem, you must follow the procedure used in class:
    define entry connector, use a wrapper W-1 to get the correct frame,
    create wrapper W-2 that extracts list of country names and URLs,
    and learn wrapper W-3 that takes as input the URLs from the previous
    one and extracts info of interest.

What to turn in:

For each learned wrapper (note that for Factbook you'll learn several wrappers) you should print four screen shots:

• one that shows that structure of the extracted data;
  examle
• one that shows the training pages (remember that you are not allowed
  to train your wrapper on the 5 pages that are used for testing);
  examle
• one that shows the learned rules;
  examle
• one that shows the output for the five pages;
  examle