test results for the estimation of 6 parameters in the heat shock model (nonlinear least-squares).

The table shows the point estimates (13) and interval estimates (14) of the measurement noise variances corresponding to the parameter set computed using a nonlinear least-squares fitting directly on the data points. The fitting was carried out with the Levenberg-Marquardt algorithm (LM). The optimization took 115 iterations (about 21 minutes running time). The interval estimates do not contain the corresponding real variances, indicating that the parameter set is invalidated by the test.