PointsUpdatemp {adlift} | R Documentation |
This function performs the update lifting step using a given configuration of neighbours and boundary handling.
PointsUpdatemp(X, coeff, nbrs, newnbrs, index, remove, pointsin, weights, lengths, updateboundhandl)
X |
the vector of grid values. |
coeff |
the vector of detail and scaling coefficients at that step of the transform. |
nbrs |
the indices (into X) of the neighbours to be used in the lifting step. |
newnbrs |
as nbrs, but repeated according to the multiple point structure of the grid. |
index |
the indices into pointsin of nbrs, the neighbours of remove. |
remove |
the index (into X) of the point to be removed. |
pointsin |
The indices of gridpoints still to be removed. |
weights |
the prediction weights obtained from the regression in the prediction step of the transform. |
lengths |
the vector of interval lengths at the present step of the transform (to be updated). |
updateboundhandl |
boundary handling in the update step. Possible values are "reflect" , "stop" and "add" . If the point to be removed is at the boundary, "reflect" updates the neighbour interval to be symmetrical about its gridpoint; "stop" extends its length up until the boundary gridpoint; and "add" increases its interval length by the interval length associated to the removed boundary point. |
The procedure performs a minimum norm update lifting step. Firstly the interval lengths are updated using the coefficients obtained. Secondly, the scaling and detail coefficient list is modified using the new interval lengths.
coeff |
vector of (modified) detail and scaling coefficients to be used in the next step of the transform. |
lengths |
the vector of interval lengths after the update step of the transform. |
r |
the index into pointsin of remove. |
N |
length(pointsin). |
weights |
The regression coefficients used in prediction. |
alpha |
the update weights used to update lengths and coeff. |
Matt Nunes (matt.nunes@bristol.ac.uk), Marina Popa (Marina.Popa@bristol.ac.uk)
AdaptNeighmp
, AdaptPredmp
, CubicPredmp
, fwtnpmp
, LinearPredmp
,
QuadPredmp
, UndoPointsUpdatemp
#read in data with multiple values... data(motorcycledata) times<-motorcycledata$time accel<-motorcycledata$accel short<-adjustx(times,accel,"mean") X<-short$sepx coeff<-short$sepx g<-short$g coefflist<-list() for (i in 1:length(g)){ coefflist[[i]]<-accel[g[[i]]] } I<-intervals(X,"reflect") l<-lengthintervals(X,I$intervals,neighbours=2,closest=TRUE) lengths<-l$lengths #work out neighbours of point to be removed (31) out<-getnbrs(X,31,order(X),2,TRUE) nbrs<-out$n nbrs newnbrs<-NULL for (i in 1:length(nbrs)){ newnbrs<-c(newnbrs,rep(nbrs[i],times=length(g[[nbrs[i]]]))) } #work out repeated neighbours using g... newnbrs p<-AdaptNeighmp(order(X),X,coefflist,coeff,nbrs,newnbrs,31,TRUE,2,"ave",g) nbrs<-p$newinfo[[3]] nbrs newnbrs<-NULL for (i in 1:length(nbrs)){ newnbrs<-c(newnbrs,rep(nbrs[i],times=length(g[[nbrs[i]]]))) } newnbrs coefflist[[31]]<-p$results[[6]][31] u<-PointsUpdatemp(X,coefflist,p$newinfo[[3]],newnbrs,p$newinfo[[4]],31,order(X),p$results[[4]] ,lengths,"add") # #and here is the updated coefficient list... u$coeff