Table 1.
Summary of investigated queries, with genuine examples in Swedish (one complete query per line) in addition to their English translations.
Figure 1.
A flow-chart of the statistical analysis.
Figure 2.
An overview of the sentinel and the laboratory data for the two investigated influenza seasons (2005/2006 and 2006/2007).
Figure 3.
The number of queries matching the selected query types plotted over time.
Table 2.
R2 values and mean predictive errors using leave-one-out cross validation for all generated models for the sentinel and the laboratory models respectively.
Figure 4.
Observed values (black), predicted values from the full models (red, with circles), and predicted values using a model fitted on data from the opposite season (blue, with squares) for the model predicting sentinel values and for the model predicting laboratory values.
Figure 5.
The relative contribution of the scaled and centred queries to each component in the model predicting sentinel values.
Figure 6.
The relative contribution of the scaled and centred queries to each component in the model predicting laboratory values.