Selective Clustering
This is one of the strongest features of NaviFind TM.
As an alternative to the category tree, this feature selects
those categories that are most likely to be relevant and
effective for any search scenario. To save time, the
Selective Clustering categories may be presented instead
of the full category tree. In most cases, these will
guide the users to the information desired with the least
number of interactions.
What guides cluster selection?
-
Information-Specific: Analyzing the information
at hand at every stage and how it is
distributed among the different categories.
-
Business Rules: Being guided by the site
owner's own considerations
(for example, the DVD library may set a low priority for presenting price ranges,
may prefer to present top level categories first, etc.)
-
Personalization: Based on the past interaction
of the user with the system,
NaviFind TM can arrange information to suit personal preferences.
-
Collaborative Filtering: Based on
the past interaction of other users doing
similar searches. This sets up group behavior patterns.
-
Segmentation: Also based on a comparison
to other users, but with the
automatic segmenting of users to interest groups and behavior groups.
System Architecture