Section 3n

3.13 Robust Regression: 3.13.1 M-Estimates, 3.13.2 Estimates Based on Robust Location and Scale Measures, 3.13.3 Measuring Robustness, 3.13.4 Other Robust Estimates

[3/100%]
Show
LaTeX
Problem 3n.1

Let χ(z)=sign⁡(∣z∣−1/c)\chi (z) = \operatorname {sign}(\left|z\right|-1 / c) for some constant cc. Show that the solution of (3.77) is the MAD estimate

s=cmedian⁡i∣ei(b)∣ s = c \operatorname {median}_{i}\left|e_{i}(\mathbf{b})\right|
?
Problem 3n.2

Show that if we put wij=1w_{i j} = 1 in (3.99), we get Jaeckel's estimator defined by (3.97) with Wilcoxon weights.

?
Problem 3n.3

Show that if ss is the solution of (3.101) with ρ\rho given by (3.102), then the resulting ss is approximately the lower quartile of the (n2)\binom {n}{2} differences ∣ei(b)−ej(b)∣\left|e_{i}(\mathbf{b})-e_{j}(\mathbf{b})\right|.

?