Data Mining for Personal Navigation

levelsordΔιαχείριση Δεδομένων

20 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

86 εμφανίσεις

Data Mining for Personal
Navigation

Gurushyam Hariharan

Pasi Fränti

Sandeep Mehta


DYNAMAP PROJECT

University of Joensuu, FINLAND

http://cs.joensuu.fi/pages/franti/dynamap
/


Personal Navigation


Location information is used for:


1.
Plotting location of user on a map

2.
Navigational guidance to given
destination

3.
Provide data related to location


Data mining required


1.
For retrieval of location
-
related
information from www (
Web mining
)

2.
For Task
-
oriented data extraction from
web documents

3.
For user profiling (additional
parameters for defining what is
relevant)


Overall Scheme





Traditional
definition of r
elevance



Keyword


Web
-
Based Search Engines


User Profile


Past Behavior of self and community define
the profile


Automatic suggestions
(e.
g. Amazon.com
proposed other
“relevant” books)




Novel Approach to Data(Web)
-
Mining for a MOBILE USER


Key is to find
RELEVANT
information


Re
-
defining Relevance for Mining Web


Relevance depends on



User request at the moment



User preferences


Relevance = Traditional Parameters



(Keywords, Profile) + LOCATION



Additional relevance factor:
Location


Co
-
ordinates
of
mobile User


City/Street
address


Relevance

=

Location

+
Keywords
(+
Profile
)


For example:


Helsink
i downtown


“Restaurant”



Budget prize
” “
Vegetarian




Issues for a Personal Navigation
System with the NEW Definition


Spot the client on the Globe


Co
-
ordinate


Location interconvertion


Data Extraction: Task oriented search of web


Scalability (in accordance with User’s Mobile
Device)


User profile learning


Pass Relevant Information to the Mobile User

Possible use scenario


Scenarion including user profiling



WE REGRET...

... the absence of the authors.


Gurushyam did not get VISA to USA


Pasi is busy elsewhere and could not
change his plans in such short notice:

T
HANK YOU

!


For more information, contact:



franti@cs.joensuu.fi


http://cs.joensuu.fi/pages/franti/dynamap/