Skip to main content
. 2016 Jun 24;16(7):958. doi: 10.3390/s16070958

Table 4.

Predictive accuracy for the generated maps (higher is better).

Model 1 PFA DT PNN KNN RProp NB
Walk 0.977 0.925 0.924 0.925 0.452 0.504
Rnd_Walk 0.276 0.276 0.275 0.276 0.270 0.253
Crash 0.571 0.629 0.629 0.626 0.561 0.336
Rnd_Crash 0.388 0.398 0.392 0.394 0.381 0.381
Zigzag 0.485 0.478 0.478 0.475 0.454 0.456
Rnd_Zigzag 0.485 0.467 0.468 0.464 0.442 0.447
Average 0.530 0.529 0.528 0.527 0.427 0.396
Model 2 PFA DT PNN KNN RProp NB
Walk 0.288 0.516 0.079 0.124 0.448 0.078
Rnd_Walk 0.270 0.276 0.269 0.272 0.267 0.267
Crash 0.509 0.612 0.570 0.578 0.547 0.328
Rnd_Crash 0.391 0.419 0.411 0.421 0.392 0.370
Zigzag 0.522 0.554 0.556 0.543 0.504 0.460
Rnd_Zigzag 0.528 0.584 0.558 0.574 0.513 0.451
Average 0.418 0.493 0.407 0.419 0.445 0.326
Model 3 PFA DT PNN KNN RProp NB
Walk 0.288 0.079 0.080 0.119 0.080 0.071
Rnd_Walk 0.267 0.276 0.267 0.271 0.278 0.271
Crash 0.702 0.815 0.705 0.725 0.758 0.563
Rnd_Crash 0.923 0.945 0.909 0.929 0.930 0.929
Zigzag 0.881 0.960 0.885 0.910 0.871 0.836
Rnd_Zigzag 0.880 0.961 0.882 0.930 0.876 0.815
Average 0.657 0.673 0.621 0.647 0.632 0.581