Characteristic Trajectories Detection in Spatio-Temporal Data Streams

Abstract

This article represents the approach for solving task of identifying objects' frequent migration trajectories. This task may take place in the study of migration trajectories of birds, animals; in the study of the trajectories of an object or a group of objects during a period of time. The described method is oriented to processing of spatio-temporal data streams. Two stages for identifying characteristic trajectories are considered. The first step is to solve the clustering problem for the analyzed data. The second stage is identifying the most frequently encountered sets of clusters. The time period of information to be processed should be chosen. Ways and algorithms to solve these tasks are offered which take into consideration requirements for speed and quality of problem solving.