Predicting Cellular Link Failures to Improve User Experience on Smartphones

Abstract

Cellular link failures cause poor user experience perceived as call drop or slowness/stall in data transfers. In this paper, we investigate a variety of supervised models that can predict link loss events. Based on this, the device could take proactive actions to avoid the link loss and improve the user experience in a seamless way. Features are chosen to represent wireless link performance and are averaged to mitigate temporal variations. We also introduce methods to improve precision while trading off with recall.

Extracted Key Phrases

Cite this paper

@inproceedings{TomPredictingCL, title={Predicting Cellular Link Failures to Improve User Experience on Smartphones}, author={Alex T Tom and S. Vasudevan} }