Logistic regression was used to analyse the overall omissions between the two groups.
We evaluated algorithm performance by overall accuracy, omission error, and commission error for both cloud and cloud shadow.
Globally, the standard errors were reduced by 63%, 54%, 22% and 53% for overall accuracy, omission error, commission error and total burned area estimates respectively.
We also performed sensitivity analyses to determine the effects of omitting a single study on the overall effect; omission of a single study changed the overall OR from 5.15 (2.85, 9.29) to 6.09 (3.37, 11.0) using the random method, suggesting that there was no single sensitive study (Table 4).
We calculated the response variance explained, the area under the receiver operator characteristic curve (AUC), the overall accuracy, the omission error rate and the commission error rate based on the aggregated CV results.
Overall, 75% of omissions were lack of documentation of investigations, either radiology or pathology.
No individual study affected the overall RR predominantly, since omission of any single study did not make a large difference.
Reported measures were overall accuracy, errors of omission (EOM), errors of commission (EOC), correct response latency, standard deviation of response latency, coefficient of variation (COV) in response latency, and measures of fluctuation in attentional state [ 13].
Furthermore, the significance of lexical category (F[2,24] = 113.99; p < 0.0001) showed an overall larger number of omissions to pseudo-words than to words (words = 7.12%; pseudowords = 18.41%).
However, overall only 101 of 199 omissions detected at chart round were subsequently corrected (51% of all omissions), when checked in September 2009.
