[Robast-commits] r208 - pkg/ROptEst/inst/scripts

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

Author: stamats
Date: 2008-11-03 09:30:41 +0100 (Mon, 03 Nov 2008)
New Revision: 208

Modified:
   pkg/ROptEst/inst/scripts/UnderOverShootRisk.R
Log:
added confidence interval, seems to work correctly ...

Modified: pkg/ROptEst/inst/scripts/UnderOverShootRisk.R
===================================================================
--- pkg/ROptEst/inst/scripts/UnderOverShootRisk.R	2008-11-03 08:23:11 UTC (rev 207)
+++ pkg/ROptEst/inst/scripts/UnderOverShootRisk.R	2008-11-03 08:30:41 UTC (rev 208)
@@ -1,7 +1,7 @@
 ###############################################################################
 ## Example: Normal Location
 ###############################################################################
-system.time(require(ROptEst))
+require(ROptEst)
 
 ## generates Normal Location Family with mean = 0
 N0 <- NormLocationFamily(mean=0) 
@@ -103,11 +103,16 @@
 N0.IC7 <- optIC(model=N0.Rob5, risk=asUnOvShoot(width = 1.960))
 (Mest1 <- locMEstimator(X, IC=N0.IC7))
 
+## confidence interval
+confint(Mest1, symmetricBias())
+
 N0.Rob6 <- FixRobModel(center = NormLocationFamily(mean = 0), 
                        neighbor = ContNeighborhood(radius = 0.05))
 N0.IC8 <- optIC(model = N0.Rob6, risk=fiUnOvShoot(width = 1.960/sqrt(n)), sampleSize = 1e2)
 (Mest2 <- locMEstimator(X, IC=N0.IC8))
 
+## confidence interval
+confint(Mest2, symmetricBias())
 
 ## 3. Kolmogorov(-Smirnov) minimum distance estimator
 (est0 <- MDEstimator(x=X, NormLocationFamily()))
@@ -117,7 +122,10 @@
                        neighbor = ContNeighborhood(radius=0.5))
 N0.IC9 <- optIC(model=N0.Rob7, risk=asUnOvShoot(width = 1.960))
 (est1 <- oneStepEstimator(X, IC = N0.IC9, start = estimate(est0)))
+confint(est1, symmetricBias())
+
 N0.Rob8 <- FixRobModel(center = NormLocationFamily(mean = estimate(est0)), 
                        neighbor = ContNeighborhood(radius=0.05))
 N0.IC10 <- optIC(model=N0.Rob8, risk=fiUnOvShoot(width = 1.960/sqrt(n)), sampleSize = 1e2)
 (est2 <- oneStepEstimator(X, IC = N0.IC10, start = estimate(est0)))
+confint(est2, symmetricBias())