Skip to main content
Fig. 2 | BMC Medical Research Methodology

Fig. 2

From: Spatiotemporal variations in exposure: Chagas disease in Colombia as a case study

Fig. 2

Goodness-of-fit of the model averaging of the 3 modelling approaches for all serosurveys. The solid lines and envelopes show standardised distances between observations and predictions’ median (blue), and 95%CrI (upper bound in red and lower bound in purple). A perfect fit would translate in all colored solid lines overalpping with the correspondingly-colored dotted lines. A blue solid line overlapping the blue dotted line, together with a red and purple solid lines at 2 and − 2 respectively would reflect a good central prediction with CrI in predictions twice as large as the CrI in the ‘observed’ FoI. Approach 1: models fitted with median FoI estimates and selected based on predictive R2; Approach 2: models fitted with median FoI estimates and selected based on predictive R2 and overlap; Approach 3: models fitted with the full posterior distribution of FoI estimates and selected based on the predictive R2 and overlap

Back to article page