At the heart of any search endeavour,
no matter what kind of search tool you are using,
there are three areas that can affect your search
result significantly:
-
Content of search engine
-
Search logic or algorithm
-
Presentation of search result
We discussed last week how content of
search engines can significantly affect your search
result. This week, we look at search logic.
When confronted with a search engine,
most of us type one or few words in search window
and hope the search result shows most relevant information.
Now, what can you do besides typing a
few relevant words into the search window ?
Well, actually a lot ! You can specify
that words must be in the title of a page, or that
specifying words must be in an URL, or in a special
HTML tag or you can use logical operators between
words like AND, OR, and NOT....
Most search engines employ keyword-based
search system, that work on occurrence or non-occurrence
of search word to identify relevant documents in its
database. Though a few search engines have started
using fuzzy logic based searches - still these are
in experimental basis.
While keyword based searches are extremely
powerful and ideally suited for database search -
its major weakness is its inability to understand
word meaning. As a result, it searches mechanically
for words - leaving the problem of synonyms squarely
on searcher's shoulder.
To give an example - searching for 'garment'
will not get you documents dealing in 'apparel' or
'clothes' or 'T-Shirt'. Worse, the search engine may
even overlook the keyword 'garments' - unless it has
facility for truncation.
So, as searcher, it becomes your responsibility
to construct search expression in such a way that
it gets maximum relevant documents from search engine
database.
Every search engine provides a set of
tools for constructing efficient search expression.
Perhaps the most important of all is the Boolean search
operators.
What is Boolean
Search ?
Boolean searching is based on a system
of symbolic logic developed by George Boole, a 19th
century English mathematician. Most keyword searchable
computer databases support Boolean searches. Boolean
search techniques may be used to perform accurate
searches without producing many irrelevant documents.
When you perform a Boolean search, you
search the computer database for the keywords that
best describe your topic. The power of Boolean searching
is based on combinations of keywords with connecting
terms called operators. The three basic operators
are the terms AND, OR, and NOT.
AND
This operator combines two search words
in a search expression that retrieves documents containing
BOTH the words.
For example: The search expression 'Fabrics
AND Buyer' will retrieve all documents that have both
the words 'fabrics' and 'buyer
AND operator narrows a search. More words
you combine with AND - greater will be precision in
search result. However, you may also miss out relevant
documents that do not contain ANY OF THE words specified.
OR
The OR operator combines two search words
in a search expression that retrieves documents containing
ANY of the words.
For example, the search expression 'Agent
OR Agents OR Distributor OR Distributors' will fetch
all documents that have ANY of the search words.
Obviously, OR operator broadens or widens
a search to include documents containing any of the
search words. This operator is particularly useful
when there are several common synonyms for a concept
or variant spellings of a word. However, indiscriminate
use of OR operator may fetch junk or unwanted documents.
NOT
NOT operator excludes unwanted documents
having the specified search word. For example, the
search expression 'Buyer NOT Agent' first fetches
all documents that have the word Buyer then goes on
to remove all documents from this collection that
have the word 'agent' in it. The result is a collection
of documents that have the word buyer but not the
word agent.
Boolean search terms may be combined
in various ways to carefully refine searches. For
example:
(Buyer OR 'Buying Agent') AND (UK OR
England OR Britain)
NOTE: A phrase
(i.e more than one word) is always marked with inverted
comma (e.g. 'Buying Agent')
This place is too short for more explanation.
Interested readers may check following tutorials for
graphical illustration of Boolean operators.
