[Robast-commits] r48 - pkg/RobLox/man

[noreply at r-forge.r-project.org]

Author: stamats
Date: 2008-02-19 09:10:33 +0100 (Tue, 19 Feb 2008)
New Revision: 48

Modified:
   pkg/RobLox/man/rlsOptIC.AL.Rd
   pkg/RobLox/man/roblox.Rd
Log:
switched from ksEstimator to MDEstimator

Modified: pkg/RobLox/man/rlsOptIC.AL.Rd
===================================================================
--- pkg/RobLox/man/rlsOptIC.AL.Rd	2008-02-19 06:14:51 UTC (rev 47)
+++ pkg/RobLox/man/rlsOptIC.AL.Rd	2008-02-19 08:10:33 UTC (rev 48)
@@ -75,17 +75,17 @@
 
 ## 2. Kolmogorov(-Smirnov) minimum distance estimator
 ## -> we use it as initial estimate for one-step construction
-(est0 <- ksEstimator(x=x, Norm()))
+(est0 <- MDEstimator(x, ParamFamily = NormLocationScaleFamily(), distance = KolmogorovDist))
 
 ## 3. one-step estimation: radius known
-IC1 <- rlsOptIC.AL(r = 0.5, mean = est0$mean, sd = est0$sd)
-(est1 <- oneStepEstimator(x, IC1, est0))
+IC1 <- rlsOptIC.AL(r = 0.5, mean = est0$estimate[1], sd = est0$estimate[2])
+(est1 <- oneStepEstimator(x, IC1, est0$estimate))
 
 ## 4. one-step estimation: radius unknown
 ## take least favorable radius r = 0.579
 ## cf. Table 8.1 in Kohl(2005)
-IC2 <- rlsOptIC.AL(r = 0.579, mean = est0$mean, sd = est0$sd)
-(est2 <- oneStepEstimator(x, IC2, est0))
+IC2 <- rlsOptIC.AL(r = 0.579, mean = est0$estimate[1], sd = est0$estimate[2])
+(est2 <- oneStepEstimator(x, IC2, est0$estimate))
 }
 \concept{normal location and scale}
 \concept{influence curve}

Modified: pkg/RobLox/man/roblox.Rd
===================================================================
--- pkg/RobLox/man/roblox.Rd	2008-02-19 06:14:51 UTC (rev 47)
+++ pkg/RobLox/man/roblox.Rd	2008-02-19 08:10:33 UTC (rev 48)
@@ -115,12 +115,16 @@
 ## estimator comparison
 # classical optimal (non-robust)
 c(mean(x), sd(x))
+
 # most robust
 c(median(x), mad(x))
+
 # Kolmogorov(-Smirnov) minimum distance estimator (robust)
-ksEstimator(x, distribution = Norm())
+MDEstimator(x, ParamFamily = NormLocationScaleFamily(), distance = KolmogorovDist)
+
 # optimally robust (amount of gross errors known)
 c(res1$mean, res1$sd)
+
 # optimally robust (amount of gross errors unknown)
 c(res2$mean, res2$sd)