Table I.
Counts of true positive (TP), false positive (FP), true negative (TN) and false negative (FN) arrow placements for weeks 47–6 from the 2013–2014 flu season is shown
| Week | 47 | 48 | 49 | 50 | 51 | 52 | 1 | 2 | 3 | 4 | 5 | 6 | Average |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Correctly placed (TP) | 0 | 4 | 5 | 4 | 7 | 3 | 7 | 3 | 7 | 10 | 5 | 1 | 4.7 |
| Incorrectly placed (FP) | 3 | 2 | 4 | 5 | 0 | 0 | 3 | 17 | 11 | 9 | 21 | 13 | 7 3 |
| Correctly excluded (TN) | 45 | 41 | 39 | 39 | 39 | 23 | 31 | 28 | 25 | 24 | 22 | 30 | 31 8 |
| Incorrectly excluded (FN) | 0 | 1 | 0 | 0 | 0 | 22 | 7 | 0 | 5 | 5 | 0 | 4 | 4.2 |
| Accuracy (%) | 93.8 | 93.8 | 91. 7 | 89.6 | 100.0 | 54.2 | 79 2 | 64 6 | 66 7 | 70 8 | 56.3 | 64 6 | 76 0 |
| AUC | 0.00 | 0.88 | 0.95 | 0.94 | 1 | 0.56 | 0.71 | 0.81 | 0.64 | 0.7 | 0.76 | 0.45 | 0.67 |
| F-measure | 0.00 | 0.73 | 0.71 | 0.62 | 1.00 | 0.21 | 0.58 | 0.26 | 0.47 | 0.59 | 0.32 | 0.11 | 0.45 |
Notes: Weeks 45 and 46 are not included because we need at least two weeks of data to be able to generate predictions. Our accuracy percentages tell us how many arrows among the 48 lower states in the USA were placed correctly. In this example, the arrow placement was correct on average 76% of the time