AIRFORCE - AIR FOReCast in Europe
Overview
Background & policy context:
The complexity of passenger movements, combined with the random aspect of travelling needs, make it difficult to generate accurate models based on reasoning or expertise only. Recent breakthroughs in data mining algorithms, however, if applied to this domain, could support the identification of existing relationships between causes and effects in huge numbers of events, whereas a human being would simply be overwhelmed with information. Combined with text mining of event-related textual information on the Internet in various languages, these methods can be the seed of very innovative and efficient forecasting tools. The AIRFORCE results are of general interest because they can be used in a range of related areas and they contribute to reinforcing the competitiveness of European companies using this technology.
Objectives:
The AIRFORCE Project proposed developing and evaluating a high-performance forecasting system providing better services to European citizens in the air domain.
Methodology:
New advanced statistical methods have been combined with intelligent monitoring and interpretation of linguistic information available on the Internet to detect factors having a real influence on passengers movements and build efficient computational forecast models:
- to evaluate the contribution of advanced statistic methods to forecast the number of air passengers for various destinations in Europe;
- to develop methods allowing the automatic extraction of limited sets of influencing parameters from huge amounts of data;
- to build up a permanently refreshed database of events that will contribute to a better understanding of motivation for travel of the customers.
Share this page