130 likes | 369 Views
Parameter Estimation using Least Squares. u nknown parameters. Least Squares. s calar variables. m easured value. Least squares identification. attempts to find values for theta for which the left- and the right-hand-sides of differ by the smallest possible error. More precisely,
E N D
unknown parameters Least Squares scalar variables measured value
Least squares identification • attempts to find values fortheta for which the left- and the right-hand-sides of • differ by the smallest possible error. More precisely, • the values of theta leading to the smallest possible sum of squares for the errors over the N experiments.
Least squares model fitting observation error Linear regression observed variable unknown coef. The aim is to find the value of which minimizes the cost function
Least squares model fitting • Example: consider a temperature measuring device with a voltage output, u. It is known that the temperature, y, is a function of the output voltage, the model is given by
Least squares model fitting For N samples The value that minimisesV makes the gradient of V with respect to zero