Performance study of data stream approximation algorithms in wireless sensor networks

Reducing amount of data transmitted enables conserving scarce battery power in wireless sensor networks. In our previous work, we propose two data approximation algorithms for data reduction in sensor networks, maintaining the accuracy of query results within certain bounds. In this paper, we provide a performance study and analysis of these algorithms with emphasis on the types of data for which the algorithms are appropriate. We experimented with different data sets to determine the reduction ratios achieved , energy consumed, errors introduced, complexity of query answering obtained. We provide comparison of our algorithms with related methods. The presented results indicate the superiority of our methods in terms of data reduction and accuracy of query results.