Fig. 3: Exemplar plot of calibration and data for objective four “Multiplicity of infection”, with exemplar epidemiological predictions of prevalence vs. EIR for the final calibration, and sensitivity of fitting objectives to each parameter.
a Multiplicity of infection by age. Comparison of simulator goodness of fit for objective 4, the age-specific multiplicity of infection (number of genetically distinct parasite strains concurrently present in one host). Simulations were carried out for the same random seed for all parameterizations and for a population size of N = 5000. b Simulated epidemiological relationship between transmission intensity (entomological inoculation rate, EIR) and P. falciparum prevalence rate (PfPR2–10). Simulated epidemiological relationship between the transmission intensity (EIR in number of infectious bites per person per year) and infection prevalence in individuals aged 2–10 years (PfPR2–10) under the parameterizations achieved by the different optimization algorithms. Lines show the mean across 100 random seed simulations for a simulated population size N = 10,000 and the shaded area shows the minimum to maximum range. c Parameter effects on the objective variance. Using the GP emulator, a global sensitivity analysis (Sobol analysis) was conducted. The tile shading shows the total effect indices for all objective functions and parameters grouped by function. SEN Senegal, TZN Tanzania.
