Geoscience Reference
In-Depth Information
homologous. However, the choice of the threshold is very important. More
details were mentioned in (Karam et al. 2010).
4.1.2
For the place names' integration, the fusion technique uses the
Levenshtein distance to compare the place names (String of characters) of
two objects from two different providers (Karam et al. 2010).
4.1.3
Semantic integration between these two objects is related to their
metadata/data differences (e.g. phone, email, website, etc.). To avoid
duplication of the service details from two different providers, Levenshtein
distance algorithm had been used to check for homologous information
and the agreed results were deducted from a matching table in our frame-
work MPLoM (Multi Providers LBS on Mobile devices). A semantic on-
tology-driven approach could be implemented via Protégé as a second
more intelligent solution using inference reasoning. For example, if a
pedestrian wants to know what restaurants can offer “Hamburger”, the plat-
form should list all the restaurants of type American or Fast food.
4.1.4
We can assume that difficulties in location integration had been par-
tially solved by the above solutions. The final decision for homologous ob-
jects depends on the output result of the belief function with Dempster op-
erator. Geographic positions, place names and semantic details results are
assigned each one a certain weight, reflecting the degree of candidates'
homogeneity towards integration. Dempster operator will combine the
three different weights and as far as their sum is high, the probability to
consider both (Karam et al. 2010).
4.2 Section 2: Main Contribution for Cartographic Integration
Once the integration problems were solved at the information level, we
will consider the integration at the cartographic visual level. One should be
able to visualize on the screen a unique base map whose components are
retrieved from the various providers contrary of what is shown in
Fig. 2
and 3 below.
