Advanced Searching

1
Advanced Searching

• Use Query Languages.
• Use more than one search engine.
• Or metasearches like at www.metacrawler.com
• Start with simple searches.
• Add new parameters incrementally.

2
Query Language Features

• Find an exact string.
• Quotes around words
• around words.
• Find words near each other.
• President near Hillary
• Find words following each other.
• adams following president

3
Query Language Features

• Both words must appear
• AND between words
• before both words
• At least one word must be there
• OR between words.
• Possibly words in parentheses (movies music)
• Wildcards hits on similar or partial names.
• , ?

4
Query Language Features

• Text keyword finds information in the text of the
  document.
• Anchor keywords finds information in hypertext
  links in the document.
• Host keyword finds information only at a
  particular site.

5
Searches for official pages

• Look up the topic you want to find.
• Look at who owns the site.
• See if its a free site or a hosting site or
  whether its for a large company that owns the
  property.
• A TV show may have more than one really
  good-looking site, even if some are fansites.

6
How Search Engines Work

• Most sites use crawlers/spiders/robots
• They go to web pages and follow links
• Return pages and links
• Some directories use humans to compile
• Most sites use some combination of both.
• For example www.looksmart.com

7
Crawler Software

• Constantly scans the web.
• Looks for new and updated pages.
• Follows links to new pages.
• May look at META tags for keywords.
• Sends the data to the index/catalog.

8
Index or Catalog

• Receives data from the crawler.
• Organizes the data.
• Stores the data.
• Updates the data.
• May have human intervention.
• Most parts are automated, however.
• Used as data source for the search engine.

9
Search Engine

• Waits for queries to come in.
• Figures out what the queries mean.
• Sends data requests to the catalog.
• Retrieves responses from the catalog.
• Sends responses to the query initiator.
• May use other databases.
• USENET archive on www.google.com

10
How Pages are Ranked

• Location keywords in the page.
• Words in the title are given more weight.
• Words at the start or larger are given more
  weight.
• Frequency of the keywords.
• More frequent appearance of the words improves
  the ranking.

11
How Pages are Ranked

• Following links
• More links to a page may help rankings.
• Search engines try to stop fake links.
• A higher quality of links may help rankings.
• Clickthrough
• Clicking on a link may give it a higher ranking.
• Search engines try to stop fake clicking.

12
Paid Listings

• Some sites allow paid listings.
• www.overture.com is all paid listings.
• Most sites have some paid listings.
• www.google.com sponsored links at top/right.
• www.lycos.com has sponsored links at top.
• www.yahoo.com has sponsored links at top.
• www.looksmart.com has sponsored links at top.
• Other sites have sponsored links as well.

13
Search Results Format

• Generally, there are sponsored listings or
  affiliates listed first.
• Category-based engines list directories.
• www.yahoo.com
• www.looksmart.com
• www.excite.com lists other searches that it
  thinks you might have wanted to do instead.

14
Project 4 Hints/Other Searches

• www.mapquest.com
• www.ipl.org
• www.switchboard.com
• www.nic.gov for .gov sites.
• For the 5 webpages on your major, you must pick
  different domains for each site.
• www.smartraveler.com