A New Path Loss Predicting Strategy for Radio Network Planning

This paper presents a new clustering based strategy for path loss prediction. We demonstrate how data mining technology can be utilized to explore the knowledge existing in massive geographical data thus improving path loss prediction accuracy and reducing field measurement cost. Comparative path loss prediction experiments between new strategy and traditional ones confirm that the new strategy guarantees the high predicting accuracy by making full use of Geographical Information System (GIS), and significantly reduces the cost/time requirements.