Evaluation and comparison of prediction stability for user movement pattern detection algorithms

Abstract

With increasing availability of high bandwidth mobile networks, ranging from classical UMTS up to local IEEE 802.16 cells, efficient management for mobile resources gains additional significance. Approaches targeting at detecting repeating patterns in mobile user movements enable network operators to react to cell handovers and bottle necks in advance. This paper discusses several methods to extract movement patterns from mobile users and analyses sensitiveness to mobility data and algorithm parametrisation.