On Using Social Context to Model Information Retrieval and Collaboration in Scientific Research Community - Community-Built Databases: Research and Development - page 136

1. User behavior : user behavior data consists of user-search engine interaction
features such as click through data, eye-tracking, and browsing features. These
various interaction features constitute the dynamic context [ 11 , 12 ] about the
user's experience allowing a robust prediction to be made about his preferences
and short-term or long-term preferences, when seeking information.
2. User interests : generally, this factor expresses the cognitive background of the user
that has an impact on his relevance judgment. The great benefit of using the user
interests [ 13 , 14 ] is to disambiguate queries and improve retrieval precision from a
large amount of information.
3. Application : user's application refers to the user's background in accordance
with the principles of evidence-based domain of interest such as diagnosis in
medicine. Domain-dependent applications provide clues allowing the user's
information need to be better ascertained. The main objective of using applica-
tion-dependent features in the retrieval process is to interpret more accurately
the user's information need within a more restricted domain so as to provide
specific answers [ 15 ]. In order to achieve this goal, specific domain tasks are
identified with related specific queries and guidelines for selecting relevant
results.
4. Task : task could be defined as the goal of information seeking behavior [ 16 ].
Numerous tasks may be achieved by the users like reading news, searching for
jobs, preparing course material, and shopping. The aim of considering this factor
is to understand the purpose of user queries in order to deliver more accurate
results. In web document retrieval, user queries can achieve three main (general)
tasks: the topic relevance task, the homepage finding task, and service finding
task. Appropriate query and document features are then exploited to predict the
desired task and rerank the results. In mobile IR, task could be defined as the
application's achievement such as tourist guide or GPS 1 - based transport .
5. Location : this factor concerns the geographical zone of interest corresponding to
the current query. It is particularly used to categorize queries as local or global
ones. Compared with global queries, local ones are likely to be of interest only to
a searcher in the relatively narrow personal region. As an example, a search for
housing is a location-sensitive query. Techniques are applied [ 17 ] to identify the
specific geographical locality addressed by a query in order to improve the
quality of the query results.
By considering the personal context during document retrieval, personalized
information retrieval is achieved. The key idea behind personalizing information
retrieval is to customize search based on specific personal characteristics of the user.
Therefore, as a personalized search engine is intended for a wide variety of users with
personal contexts, it has to learn the user model first (commonly called user profile)
and then exploit it in order to tailor the retrieval task to the particular user.
1 Global positioning system.