Représentation Symbolique de Trajectoires Géolocalisées.

Lot of geolocalizated trajectories are produced a nd available but their exploitation to extract interesting information remains difficul t. The obstacle we address here is the poor quality of some of these data: very irregular sampl ing and poor precision. We propose to transpose them into symbolic representations to per form treatments such as the application of data mining methods: extraction of patterns, associ ation rules, clustering, etc. This transposition should avoid as much as possible to r ely on a priori threshold values so that the same method applies regardless of the nature of mov ing objects. In this paper, we propose to use the SAX method (Symbolic Aggregate approXimatio n) for the symbolic representation of moving objects trajectories. SAX is used to represe nt a series of one-dimensional numerical values by a series of symbolic values automatically determined in relation to the distribution of numerical values. We present two applications of SAX to moving objects trajectories