A Unified Prediction Framework for Signal Maps

Signal maps are essential for the planning and operation of cellular networks. However, the measurements needed to create such maps are expensive, often biased, not always reflecting the metrics of interest, and posing privacy risks. In this paper, we develop a unified framework for predicting cellular signal maps from limited measurements. We propose and combine three mechanisms that deal with the fact that not all measurements are equally important for a particular prediction task. First, we design \emph{quality-of-service functions ($Q$)}, including signal strength (RSRP) but also other metrics of interest, such as coverage (improving recall by 76\%-92\%) and call drop probability (reducing error by as much as 32\%). By implicitly altering the training loss function, quality functions can also improve prediction for RSRP itself where it matters (e.g. MSE reduction up to 27\% in the low signal strength regime, where errors are critical). Second, we introduce \emph{weight functions} ($W$) to specify the relative importance of prediction at different parts of the feature space. We propose re-weighting based on importance sampling to obtain unbiased estimators when the sampling and target distributions mismatch(yielding 20\% improvement for targets on spatially uniform loss or on user population density). Third, we apply the {\em Data Shapley} framework for the first time in this context: to assign values ($\phi$) to individual measurement points, which capture the importance of their contribution to the prediction task. This can improve prediction (e.g. from 64\% to 94\% in recall for coverage loss) by removing points with negative values, and can also enable data minimization (i.e. we show that we can remove 70\% of data w/o loss in performance). We evaluate our methods and demonstrate significant improvement in prediction performance, using several real-world datasets.