Analysis of the bias and variance of parameter estimates induced by different methods for the assessment of measurement noise.

<p>The figure shows the result of two hundred independent parameter estimation runs for one hundred different sets of simulated data for the parameters A, B, k, k and , where applicable. The amount of measurement noise was assessed by a pre-processing approach (a), by using a simultaneous estimation of dynamics and measurement noise (b) and using the true values for and in the estimation (c).