9

Document Sample
9 Powered By Docstoc
					                            91.113 Exploring the Internet, Fall 2011




                                             Lecture 9 Search Engine



                                                            Instructor: Jie Yang
                                                            Department of Computer Science
                                                            University of Massachusetts Lowell




Some slides courtesy off Eugene Barsky and Suite101.com
                                      2




• Google Instant
 ▫ http://youtu.be/qcm0rG8EKXI?hd=1
                                                                                                  3



Major search engines today
• Today Google, Yahoo, and MSN are most popular
  search engines, according to Nielsen statistics and
  searchenginewatch.com.
   Explicit Core Share* of U.S. Searches Among Leading Providers, September 2010 vs August 2010

                                                Share of Searches (%)
       Domain           August 2010       September 2010         Month-over-Month Point
                                                                      Change (%)

     Google Sites           65.4                66.1                        0.7

      Yahoo Sites            17.4               16.7                        -0.7

    Microsoft Sites          11.1               11.2                        0.1

     Ask Network             3.8                3.7                         -0.1

     AOL Network             2.3                2.3                         0.0


                      http://searchenginewatch.com/3641456
                                                    4



Searching
 • Search engines are powerful
 • People can use search engines without knowing
   much about using them or how they work
 • Search engines are designed to be easy to use,
   but you still need a strategy when looking for
   information
 • Some popular search engines
  ▫ Google: www.google.com
  ▫ Microsoft: www.bing.com
  ▫ Yahoo: www.yahoo.com
                                      5


Search Engines use software
called Bots or Spiders to scour the
web.
                                  6



These Bots and Spiders find new
websites and web pages by
following links on a web page.
                                  7



Once they find a web page, they
“read” the text-based content…
                                       8


…and the Search Engine stores
this data into a huge library called
an Index.
                                9



When you search for something
on Google…
                                  10



…the Search Engine reaches into
its gigantic Index…
                       11




…and before
displaying a list of
web pages it uses
its Algorithm to
calculate which
ones best match
your search query…
                                          12


…then it eventually presents you with
search results that are related to your
query…
                                      13


…which leads you to the website you
were looking for
                                                          14

How Google search engine works
(in one graph)




      http://www.codeproject.com/KB/docview/Search.aspx
                                                              15


How search engines work
• Each search engine is implemented in a unique way
  using different parameters to put webpages as first hits.
   ▫ This explains why different search engines (e.g.
     Google, Yahoo, Bing) sometimes return different
     results given the same keywords
• Google’s PageRank algorithm uses a number of links to
  a webpage as one of the major factors. (There are more
  than 100 other factors as well.)
                                          16


Searching from your browser
• More convenient than searching in the
  Google.com homepage
17
                                                  18


Basic search in Google
• How can we find out on Google whether vitamin
  C can help or prevent common cold?

                      0.35 seconds!
                                                       19


Phrase searching using quotes “”
• Putting quotes around a phrase to limit results to
  the exact phrase
                                                                        20


Boost searching experience using
Boolean logic
• What is Boolean logic?
                              Search Terms             No. of results

                             Mineral AND deposit          9,610,000




                                mineral OR ore           280,000,000
                                      or
                                 mineral | ore




                           "mineral deposit" -mining      5,490,000
                                                              21


Limit the websites you search – site:
• Search for your topic by limiting your search to reliable
  websites or domains, e.g., .gov, .edu, .org domains
                                            22


Google Advanced Search




    http://www.google.com/advanced_search
                                                        23


  Search certain file types -- filetype:
   • This limits retrieved documents by format, e.g.,
     .pdf; .doc; or .ppt




Google recognizes 13 main
file types through advanced
search, including all Microsoft
Office Document types, Lotus,
PostScript, Shockwave Flash
and plain text files.
                                                                          24

In title search – intitle:
 • Title and URL commands are among the most powerful as they
   improve our precision drastically. Very often title words convey the
   subject content of the information
 • Let’s see what the U S Food and Drug Administration (FDA) says
   about recalls




This is an example of grey
literature – reliable
documents that are hard
to approach via traditional
sources and therefore only
searchable via the Web,
by search engines
                                                           25

In URL search – inurl:
• The modifier inurl: only searches the web address of a
  page
• Let’s see what webpages on the FDA website contain
  keyword “recall” in the page names
                                                                 26

Related search operator – related:
• Retrieves documents with similar semantic patterns
• Let’s assume we want to retrieve webpages similar to the FDA
  website
                                  27

Related search operator – link:

• Find all
  websites that
  link to a
  particular
  webpage

• Let’s see who
  links to the FDA
  website
  homepage
                             28


 Wildcards (*) are useful
• Expand your
  search by
  replacing any
  single word in
  your search
  query



* food recall site:fda.gov
                                                           29

Synonym search ~
• Synonym search looks for words that mean similar
  things. Use the tilde symbol before your keyword, like
  this: ~eggplant
                                                      30

Numrange operator ..
• Search for a numerical range using the numrange
  operator. For example, search for Sony TV between
  $400 and $600 with the string Sony TV $400..$600
                                                   31




Google as a Personal Assistant
• Simply enter your calculation or conversion in
  the search box, and let Google do the rest
               32




showtimes works too!
                                                            33




More at: http://www.google.com/intl/en/help/features.html
                                                     34

A few reminders
• Common words, like I, and, then and if are
  ignored by Google. These are called “stop
  words”.
• In fact, you can combine any Boolean search
  operators, as long as your syntax is correct. An
  example: "sausage and mash" -onions
  filetype:doc
• The modifier info:site_name returns information
  about the specified page.
                                              35


Compare A with X
• Compare items with "better than" and find
  similar items with "reminds me of"
                    36



Find live webcams
                                                          37


Make Google recognize faces
• If you're doing an image search for people named Rose
  and don't want any flowers to show up, a special URL
  parameter in Google's Image search will do the trick.
• Add &imgtype=face to the end of your URL when you
  do the image search to just get images of faces
                                            38




More to search
• After Google retrieves results for your
  query, you could narrow down by
  choosing what type of results you want
                                      39




Further Reading
• 101 Google tips, tricks and hacks
 ▫ http://bit.ly/dxJgvI

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:10/27/2011
language:English
pages:39