The results when evaluating the different methods ability to find stop-locations overlapping with the ground truth.

We report 5 different error measures for each method. DBSCAN-method on WiFi data achieves the best result for Matthew’s Correlation Coefficient. One reason that the GPS-based method yields the highest precision is that mobile routers are inferred as stop-locations for the WiFi based methods, but are not reported as such in the ground truth.