Utility for Location Based Queries Sample Clauses

Utility for Location Based Queries. ‌ To demonstrate the utility of released locations, we also measured the precision and recall of kNN queries at each of the 500 timestamps in the 100 trajectories with popular M. The average results of kNN from original locations and k2NN from released locations are reported in Figure 3.8 with ϵ = 1 and δ = 0.01. In Figures 3.8a and 3.8d, we show the precision and recall with k = k2. Note that in this case precision is equal to recall. We can see that when k grows precision and recall also increase because the nearest neighbors have to be found in larger areas. PIM is consistently better than LM. Next we fixed k = 5 and varied k2. Figures 3.8b and 3.8e show the precision Precision/Recall LM PIM 10 15 20 25 k 0.6 Precision 0.5 0.4 0.3 0.2 5 10 15 20 k’ LM PIM Recall LM PIM