User's location prediction system using the filtering with correlation coefficients weight

Abstract

This paper proposes User's Location Prediction System using the Filtering with Correlation Coefficients Weight. This system heterogeneous data occurred during the process of collecting context information into homogeneous data, and improves the accuracy of clustering needed for location-awareness. Also, it applies correlation coefficients weight to extract features and predicts user's location through ARIMA time-series analysis. For the evaluation of the proposed method a test bed is constructed, tested and compared with the method without weighting. © 2011 Springer-Verlag.