search

blab / ncov-forecasting-fit

Assessing accuracy of fitness model forecasts
1 stars 2 forks source link

Reviewer 1.2 #29

Open Eabousam opened 4 months ago

Eabousam commented 4 months ago

The authors did not include any incorporation of the uncertainty of estimates into their evaluation...they could at the least do an analysis using the confidence intervals of the estimates and assessing the coverage of their intervals compared to the “truth” data

TODO: Run SVI and add posterior predictive to look at coverage. This is the only place that needs to have CIs. Supplementary figure similar to Figure 2 but look at coverage over time

Eabousam commented 3 months ago

Created coverage metric, deciding on adding noise to the figure