|
Abstract:
|
English: The main goal of the project is to develop a dynamic flight target detector from scratch
finally called PaxFinder (Passengers Finder). This goal has been evolving during the course of the internship. At the beginning, the objective was far more a generic one:
“The objective of the internship is to study the public data available from web-based
social networks so Amadeus products can be tailored to fit each individual”. The first
part was clear. The public data of the people profiles coming from the social networks
should be studied in order to be used in this project. The second part was specified to
develop a dynamic flight target detector through the analysis of passengers profiles.
Some requirements and limitations were established at the beginning of the
project in order to define and limit the scope of the project. A first requirement was to
integrate Crescando into the project in order to exploit its capacities. A second one
was to only work with profile data coming from social networks and Amadeus
repositories. For example, information like the travel history and the background of a
traveler was forbidden. Paying attention to the goal and the requirements of the project, the way to do
this dynamic flight target detector was first developing a flight profiling system and
then, a profile search system.
The tasks that the project should go through in order to reach the objectives were
the following:
· Studying which social networks contain data relevant for the planned
usage.
· Studying how data can be extracted and building tool to perform the
extractions even from social networks as Amadeus databases.
· Getting familiar with Crescando.
· Participate in the definition of the model of data and creating a new
Crescando database thanks to the existing administration tools.
· Define the parameters of a profile.
· Define a methodology to classify and compare profiles.
· Develop PaxFinder.
Any target detector has to be correctly tested. This process is composed of three
steps. The first one is an offline testing. The second one is a study with real users and
the third one is to make an online version massively working with real users data.In this project we have done the offline testing. This testing would have be done
with real data coming from Amadeus repositories and social network extractions.
However, some problems related to the data consistency into Amadeus repositories
and the social networks privacy policy did manipulate the data stored in Crescando.
Anyway, the study of the most important social networks has been done, as well as
their respective data extraction tools. The documentation of this work can be read in
the annexes attached at the end of this document. |
