A simulated dataset containing 6000 independent observations on antibody titres and the corresponding infection status. The data was simulated to resemble real influenza infection and haemagglutinin + neuraminidase titre data.
sclr_two_titre_dataA data frame with 6000 observations and 6 variables:
haemagglutinin-inhibiting (HI) titre. True simulated titre on a log scale.
HI censored (observed) titre. The titre value on a log scale that would be observed in a real dataset with a typical HI assay.
Midpoint of the interval (on a log scale) to which observed HI values are censored.
neuraminidase-inhibiting titre. True simulated titre on a log scale.
haemagglutinin-inhibiting censored (observed) titre. The titre value that would be observed in a real dataset with a typical NI assay.
influenza infection status. 1 - infected. 0 - not infected
The model behind the simulation was
$$\lambda * (1 - f(\beta_0 + \beta_1 * HI + \beta_2 * NI))$$
Where
\(f\) - Inverse logit function
\(\lambda\) = 0.25
\(\beta_0\) = -7.5
\(\beta_1\) = 2
\(\beta_2\) = 2