Architecture for Including Personalization in a Mobile GIS via Semantic Web Techniques - Geographical Information Systems: Trends and Technologies - page 206

where the user is moving to. Other applications may also offer navigation
functionalities, able to guide the user from one point to another.
Tourism is one of the most popular sectors of applicability of
geographical applications and some tourism applications are able to
(Gavalas et al. 2013): 1) find out where the user is; 2) guide the user to a
point of interest; 3) identify what the user is looking at; and 4) provide some
information about the point of interest the user is looking at. Therefore, with
these functionalities smartphones become an up to date guide wherever
the users go and true Recommender Systems (RS).
And, how do these geographical applications work? There are several
categories of RS (Gavalas et al. 2013, and references therein), regarding
their target applications, the knowledge used, the way they formulate
recommendations and the algorithms they implement: collaborative
fi ltering; content-based fi ltering; knowledge-based fi ltering; demographic
fi ltering; matrix factorization; hybrid RSs.
The approach presented in this work is a hybrid RD that uses primarily
knowledge-base fi ltering, but corrected by content-based fi ltering. According
to the classifi cation from Burke (Burke 2002), the system presented here is
weighted in a fi rst approximation and switched depending on the current
situation.
Regarding the performance, many of actual approximations take
advantage of the capacities of the smartphone (see the review from Gavalas et
al. 2013), since they use the GPS and/or the network. However, when the task
to perform is costly (computationally or by its memory requirements), these
applications tend to delegate such task to a remote server (see for example
Noguera et al. 2013). Examples of these delegated tasks are the download
of the map of the zone, the retrieval of information of the points of interest
the user is looking at, or the calculus of a route from one point to another,
or some extra information which is computationally very costly. Therefore,
most geographical applications need a cellular network to work properly.
Unfortunately, the constant need for a Internet connection poses several
drawbacks that make the application not usable in all conditions:

• 3G coverage is low in most of the territory: although 3G covers the most
populated areas and usually covers 90% of the population, isolated areas
are not covered. 1 This would make geographical applications useless in
isolated zones, despite the interest tourists may have in these areas.

• Battery consumption: batteries of smartphones last only for few hours
when using the cellular network connection, making applications
useless for longer routes.

1

See for example de zones covered by Orange in Spain: most of the main cities are covered,
but not small villages or the country. http://movil.orange.es/cobertura-orange/mapa-de-
cobertura.html