Table 1 Each step of our classification pipeline (“Vaccine intent classifier”) improves both our correlation with CDC state vaccination rates and our coverage of vaccine intent users

From: Measuring vaccination coverage and concerns of vaccine holdouts from web search logs

Pipeline step

Correlation with CDC

Num vaccine intent users

Only queries

0.62

3.18M

+manual URLs

0.80

4.95M

+manual and GNN URLs

0.86

7.45M