Skip to content

Robustness under noisy data

M = 100, N = 1000, NOISE_STD = 1.0

True OE parameters

parameter value
$b_1$ 1.00000000
$f_1$ -2.40000000
$f_2$ 1.91000000
$f_3$ -0.50400000

OE-structure estimates (SM, OE, and Sippy)

method parameter mean std
SM $b_1$ 1.00050556 0.00297019
SM $f_1$ -2.39965905 0.00177515
SM $f_2$ 1.90937438 0.00329268
SM $f_3$ -0.50371130 0.00153745
OE $b_1$ 1.00046374 0.00297773
OE $f_1$ -2.39968286 0.00177458
OE $f_2$ 1.90941797 0.00329087
OE $f_3$ -0.50373130 0.00153630
Sippy $b_1$ 1.04393391 0.05858222
Sippy $f_1$ -2.37310170 0.03420298
Sippy $f_2$ 1.86040523 0.06301047
Sippy $f_3$ -0.48102405 0.02916139

Simulation errors

method mean error std error success rate (< 5%)
SM 1.046883e-03 3.883970e-04 100.0\%
OE 1.046365e-03 3.891422e-04 100.0\%
Sippy 9.714697e-03 8.968038e-03 100.0\%