The FFQ was administered to subjects twice during a 2-week period, and evaluated using the mean values of three 24-hour recalls.
Classifier assessment was carried out using the mean of area under curve (AUC), precision, modified precision, and recall values computed from the validation set.
To mitigate against this we have used the mean of 3 24-hour recalls, thus providing more robust data from this method, which may be more appropriate to use cross-culturally and with low-income participants [ 55] as well as with those who may be cautious about engaging in a study or for whom literacy is an issue [ 56].
We used the mean of three 24-h recalls to evaluate dietary nutrient intake and found a significant graded trend in total PUFAs, linoleic acid, and DHA intake, which was the lowest in the GDM group and highest in the control subjects, with the impaired GCT non-GDM group falling in the middle (P < 0.05 for all) (Table 4).
The mean recall was 98.1%.
The mean recall scores for the four word lists were 12.0, 14.2, 12.6, 12.6 respectively.
When using the 2-week recall period, the mean fish tissue Hg level of consumed fish was approximately 80 and 65 ppb for the Japanese population (second and third visits, respectively).
Results are reported using the precision and recall metrics and their harmonic mean, the F1-score.
To more quantitatively evaluate the improvement in the performance of our algorithm for both recall and precision together, we used the harmonic mean (F) of the recall and precision values for each condition (i.e. each library and threshold combination).
They used the arithmetic mean of precision and recall to measure interjudge similarity between each unique subject pair.
