Databases Reference
In-Depth Information
categorized as nouns, verbs, adverbs, adjectives, prepositions, or pronouns. Research
has shown that most, although not all, queries are composed of primarily nouns [
39
]
on stop words (a.k.a., skip words).
The user intent [
40
,
41
,
42
] is somewhat unique to Web searching. So, given the
importance and application of intent, we cover it here.
What can we learn from the underlying intent of a searcher's key terms?
Naturally, different queries often have different underlying needs, and these under-
lying needs often induce different types of searcher behavior, from click-through
behavior to browsing behavior. For example, empirical studies have noted that broad
informational queries (e.g.,
digital camera
) require more browsing by searchers rel-
ative to more focused queries (e.g.,
find a
Nokia camera
) [
43
].
The research into user intent in Web search begins with Broder [
40
], who proposed
three broad user-intent classifications for Web queries:
navigational
,
informational
,
and
transactional
. This framework was based on empirical observation, and it has
been supported by a string of empirical research in the area of Web searching. For
example, Spink and Jansen [
44
] report that e-commerce-related queries varied from
approximately 12 percent to 24 percent using various Web search engine transac-
tion logs. Jansen, Spink, and Pedersen [
45
] stated that there was a significant use
of search engines as a navigation appliance. The researchers report that the top if-
teen queries from a 2002 AltaVista search log (e.g.,
google
,
yahoo
,
ebay
,
yahoo.com
,
hotmail
,
hotmail.com
,
thumbzilla
,
www.yahoo.com
,
babelfish,
,
mapquest
,
nfl.com,
,
nl
,
weather
,
www.hotmail.com
,
and
google.com
) were all likely expressions of a navi-
gational intent. It is apparent that the hypermedia environment of the Web provides
a unique capability of using search as a specialized form of browsing. Expanding on
Broder's classifications, Rose and Levinson [
42
] classified search queries using the
categories of
informational
,
navigational
, and
resource
, with hierarchical subcatego-
ries of each.
What is user intent?
User intent is
the resource specified by the affective, cognitive, or situational goal
expressed in an interaction with a Web search engine
. Referring to Belkin's states
of a searching episode [
46
], intent is akin to goal, and expression akin to method of
interaction. Unlike goal, however, intent is concerned with how the goal is expressed
because the expression determines what type of resource the user desires to address
his or her overall goal. Pirolli [
27
, p. 65] makes a similar delineation between task
(i.e., something external) and need (i.e., the concept that drives the information-for-
aging behavior). Saracevic's stratified model [
47
,
48
] proposes that user expressions
to an information-searching system are based on affective, cognitive, or situational
strata.
Certainly, the query is a key component of this expression of intent. The importance
of the query is obvious by the considerable amount of research examining various
aspects of query formulation, reformulation, and processing [
49
,
50
,
51
,
52
]. Pirolli
[
27
, p. 65] refers to the query also as an external representation of the need. Note that
the query is often an inexact representation of the underlying intent [
53
,
54
,
55
,
56
].
Derived from research [
40
,
42
], user intent within each category is defined as one
of the following:
