diff --git a/docs/articles/nb_igraph.html b/docs/articles/nb_igraph.html index 7ab39dd..a200616 100644 --- a/docs/articles/nb_igraph.html +++ b/docs/articles/nb_igraph.html @@ -162,7 +162,7 @@ 

## Loading required package: spData
## Loading required package: Matrix
## Loading required package: sf
-
## Linking to GEOS 3.12.1, GDAL 3.9.0beta1, PROJ 9.4.0; sf_use_s2() is TRUE
+
## Linking to GEOS 3.12.1, GDAL 3.9.0, PROJ 9.4.0; sf_use_s2() is TRUE

Getting some data @@ -181,7 +181,7 @@

Getting some datasf_extSoftVersion() }

##           GEOS           GDAL         proj.4 GDAL_with_GEOS     USE_PROJ_H 
-##       "3.12.1"   "3.9.0beta1"        "9.4.0"         "true"         "true" 
+##       "3.12.1"        "3.9.0"        "9.4.0"         "true"         "true" 
 ##           PROJ 
 ##        "9.4.0"
@@ -312,13 +312,16 @@ 
Symmetric sparse matrices
 
 nb_B1 <- spdep::mat2listw(as(as(B, "TsparseMatrix"), "CsparseMatrix"),
-    style="B", zero.policy=TRUE)
-nb_B1$style

-
## [1] "B"

+    style="B", zero.policy=TRUE)
+
## Warning in sn2listw(df, style = style, zero.policy = zero.policy, from_mat2listw = TRUE): no-neighbour observations found, set zero.policy to TRUE;
+## this warning will soon become an error
+nb_B1$style
+
## [1] "B"
+
 all.equal(nb_B1$neighbours, col2, check.attributes=FALSE)
## [1] TRUE
-
+
 all.equal(attr(nb_B1$neighbours, "region.id"), attr(nb_B$neighbours, "region.id"))

 
## [1] TRUE
@@ -331,7 +334,7 @@

Log deter the eigenvalue approach with for example spatialreg::eigenw is limited by the need to operate on dense matrices in memory to solve the eigenproblem:

-
+
 rho <- 0.1
 do_spatialreg <- FALSE
 if (requireNamespace("spatialreg", quietly=TRUE)) do_spatialreg <- TRUE
@@ -346,14 +349,14 @@ 
Log deter
 "dsTMatrix", from a "ddiMatrix". The value of
 the log determinant follows, calling a sparse Cholesky decomposition
 internally for suitable input matrices.

-
+
 n <- nrow(B)
 I <- Diagonal(n)
 class(I - rho * B)

 
## [1] "dgCMatrix"
 ## attr(,"package")
 ## [1] "Matrix"

-
+
 c(determinant(I - rho * B, logarithm=TRUE)$modulus)

 
## [1] -1.44787

 
The computation of a sparse Cholesky decomposition for each value of
@@ -361,7 +364,7 @@ 
Log deter
 avoided by updating a pre-computed object; this approach provides fast
 and accurate log determinants for larger (but not very large) data
 sets:

-
+
 nW <- -B
 nChol <- Cholesky(nW, Imult=8)
 n * log(rho) + (2 * c(determinant(update(nChol, nW, 1/rho), sqrt=TRUE)$modulus))

@@ -373,7 +376,7 @@ 
Asymmetric sparse matricesThe use of row-standardisation leads to asymmetry even if the
 underlying neighbours are symmetric, unless all entities have matching
 numbers of neighbours (for example a regular grid on a torus):

-
+
 nb_W <- spdep::nb2listw(col2, style="W", zero.policy=TRUE)
 W <- as(nb_W, "CsparseMatrix")
 str(W)

@@ -386,7 +389,7 @@ 
Asymmetric sparse matrices##   .. ..$ : chr [1:49] "1" "2" "3" "4" ...
 ##   ..@ x       : num [1:230] 0.333 0.25 0.5 0.25 0.25 ...
 ##   ..@ factors : list()

-
+
 all(W == t(W))

 
## [1] FALSE

 
The lag method for listw objects is often
@@ -395,17 +398,17 @@ 
Asymmetric sparse matriceszero.policy to
 TRUE, the spatial lag of entity 21, which has no
 neighbours, is zero by construction:

-
+
 set.seed(1)
 x <- runif(n)
 r1 <- as.numeric(W %*% x)
 r2 <- lag(nb_W, x, zero.policy=TRUE)
 all.equal(r1, r2, check.attributes=FALSE)

 
## [1] TRUE

-
+
 plot(x, r1, ylim=c(0,1))

 

-
+
 c(x[21], r1[21])

 
## [1] 0.9347052 0.0000000

 

@@ -419,20 +422,20 @@ 
Log dete
 result may be a complex vector (here it is not, as discussed below). The
 appropriate determinant method for "dgCMatrix"
 objects uses an LU decomposition internally:

-
+
 rho <- 0.5
 sum(log(1 - rho * eigenw(nb_W)))

 
## [1] -1.594376

-
+
 class(I - rho * W)

 
## [1] "dgCMatrix"
 ## attr(,"package")
 ## [1] "Matrix"

-
+
 c(determinant(I - rho * W, logarithm=TRUE)$modulus)

 
## [1] -1.594376

 
We can show the internal workings of the method as:

-
+
 LU <- lu(I - rho * W)
 sum(log(abs(diag(slot(LU, "U")))))

 
## [1] -1.594376

@@ -446,7 +449,7 @@ 
Log determinants: symmetric by
 less costly numerical methods. The "W" style used the
 cardinalities of neighbour sets (row sums) to introduce row
 standardisation, and they are stored as an attribute:

-
+
 d <- attr(nb_W$weights, "comp")$d
 all.equal(d, spdep::card(col2))

 
## [1] TRUE

@@ -454,24 +457,24 @@ 
Log determinants: symmetric by
 (which must be symmetric), we can pre- and post-multiply by the square
 roots of the inverted neighbour counts, yielding a symmetric matrix with
 the appropriate properties:

-
+
 dW <- Diagonal(n, d) %*% W
 all(dW == t(dW))

 
## [1] TRUE

-
+
 isd <- Diagonal(n, 1/sqrt(d))
 isd[21,21]

 
## [1] Inf

-
+
 Ws <- as(isd %*% dW %*% isd, "symmetricMatrix")
 rowSums(Ws)[21]

 
## [1] 0

-
+
 class(Ws)

 
## [1] "dsCMatrix"
 ## attr(,"package")
 ## [1] "Matrix"

-
+
 c(determinant(I - rho * Ws, logarithm=TRUE)$modulus)

 
## [1] -1.594376

 
As can be seen, the division by the square root of zero for entity 21
@@ -492,19 +495,19 @@ 
Using eigsn is somewhat larger, use may be made of the
 eigs function in RSpectra:

-
+
 1/range(spatialreg::eigenw(nb_B))

 
## [1] -0.3212551  0.1638329

-
+
 if (!require("RSpectra", quietly=TRUE)) dothis <- FALSE

-
+
 1/c(eigs(B, k=1, which="SR")$values, eigs(B, k=1, which="LR")$values)

 
## [1] -0.3212551  0.1638329

 
In this case, the results are trivial with small k.

-
+
 1/range(eigenw(nb_W))

 
## [1] -1.544645  1.000000

-
+
 1/Re(c(eigs(W, k=1, which="SR")$values, eigs(W, k=1, which="LR")$values))

 
## [1] -1.544645  1.000000

 
Using row-standardisation has the nice feature of setting the upper
@@ -520,15 +523,15 @@ 
Converting from sym
 

 
First we’ll see how to get from sparse matrices to graphs. The mode
 of a symmetric matrix is "undirected" by definition:

-
+
 class(B)

 
## [1] "dgCMatrix"
 ## attr(,"package")
 ## [1] "Matrix"

-
+
 
## 10824 bytes

-
+
 if (!require("igraph", quietly=FALSE)) dothis <- FALSE

 
## Loading required package: igraph

 
## 
@@ -539,11 +542,11 @@ 
Converting from sym
 
## The following object is masked from 'package:base':
 ## 
 ##     union

-
+
 g1 <- graph_from_adjacency_matrix(B, mode="undirected")
 class(g1)

 
## [1] "igraph"

-
+
 
## 6544 bytes

 

@@ -554,7 +557,7 @@ 
Converting from gra
 get.adjacency chooses a particular class of sparse matrix
 to be returned, so that the conversion process typically leads many
 matrices to fewer graph types, and back to fewer matrix types:

-
+
 # Matrix 1.4-2 vulnerability work-around
 ow <- options("warn")$warn
 options("warn"=2L)
@@ -563,13 +566,13 @@ 
Converting from gra
 
## [1] "dgCMatrix"
 ## attr(,"package")
 ## [1] "Matrix"

-
+
 if (!inherits(B1, "try-error")) print(object.size(B1))

 
## 10824 bytes

-
+
 if (!inherits(B1, "try-error")) print(all.equal(B, as(B1, "CsparseMatrix")))

 
## [1] TRUE

-
+
 options("warn"=ow)

 

 

@@ -580,7 +583,7 @@ 
Graph components in spdepn.comp.nb from the early days of the package. It is useful
 to know whether an nb object is divided up into separate
 subgraphs, and which entities are members of which such subgraph.

-
+
 res <- spdep::n.comp.nb(col2)
 table(res$comp.id)

 
## 
@@ -592,18 +595,18 @@ 
Graph components in igraph<
 

 
The same result can be obtained using the clusters
 function in igraph:

-
+
 c1 <- components(g1)
 c1$no == res$nc

 
## [1] TRUE

-
+
 all.equal(c1$membership, res$comp.id)

 
## [1] "names for target but not for current"

-
+
 all.equal(c1$csize, c(table(res$comp.id)), check.attributes=FALSE)

 
## [1] TRUE

 
The same holds for the row-standardised variant:

-
+
 W <- as(spdep::nb2listw(col2, style="W", zero.policy=TRUE), "CsparseMatrix")
 g1W <- graph_from_adjacency_matrix(W, mode="directed", weighted="W")
 c1W <- components(g1W)
@@ -622,14 +625,14 @@ 
Shortest paths in weights mat
 above. The diameter measure is then the diameter of the largest
 component subgraph. Note that this generates an n x
 n matrix:

-
+
 
## [1] FALSE

-
+
 dg1 <- diameter(g1)
 dg1

 
## [1] 7

-
+
 sp_mat <- distances(g1)
 str(sp_mat)

 
##  num [1:49, 1:49] 0 1 1 2 2 3 4 3 3 4 ...
@@ -645,7 +648,7 @@ 
Shortest paths in weights matr
 in advance (the largest lag order for which the number of links is
 greater than zero), we run into the problem of how to represent missing
 neighbour information.

-
+
 nbl10 <- spdep::nblag(col2, maxlag=10)
 vals <- sapply(nbl10, function(x) sum(spdep::card(x)))
 zero <- which(vals == 0)
@@ -657,7 +660,7 @@ 
Shortest paths in weights matr
 produced by shortest.paths, we need to set all these
 non-structural zeros to infinity (the length of the path between
 unconnected nodes), and re-instate structural zeros on the diagonal:

-
+
 lmat <- lapply(nbl10[1:(zero[1]-1)], spdep::nb2mat, style="B", zero.policy=TRUE)
 mat <- matrix(0, n, n)
 for (i in seq(along=lmat)) mat = mat + i*lmat[[i]]
@@ -691,23 +694,23 @@ 
Smirnov/Anselin (2009) cyclical m
 this for each block/subgraph by testing the condition until it meets
 w[j,k] > 0, at which point it breaks. Smirnov and Anselin (2009)
 state that rook neighbours on a regular grid meet the condition:

-
+
 nb_r <- spdep::cell2nb(7, 7, type="rook")
 nb_rW <- spdep::nb2listw(nb_r, style="W")
 spdep:::find_q1_q2(nb_rW)

 
## [1] 1 1

 
One block/graph component is found, and this one meets the cyclical
 matrix condition, as also shown by the domain:

-
+
 
## [1] -1  1

 
This does not apply to the spatial weights we have been using above,
 with two non-singleton components, neither meeting the cyclical matrix
 condition:

-
+
 spdep:::find_q1_q2(nb_W)

 
## [1] 2 0

-
+
 
## [1] -1.544645  1.000000

 
By construction, all two-node connected graph components also meet
diff --git a/docs/articles/sids_models.html b/docs/articles/sids_models.html
index d4b78d5..a7c7117 100644
--- a/docs/articles/sids_models.html
+++ b/docs/articles/sids_models.html
@@ -121,7 +121,7 @@ 
Getting the data into R
 
We will be using the spdep and
 spatialreg packages, here version: spdep, version
-1.3-4, 2024-03-27, the sf package and the
+1.3-4, 2024-05-31, the sf package and the
 tmap package. The data from the sources referred to
 above is documented in the help
 page for the nc.sids data set in
diff --git a/docs/news/index.html b/docs/news/index.html
index bb534c0..728e711 100644
--- a/docs/news/index.html
+++ b/docs/news/index.html
@@ -64,8 +64,11 @@ 
Changelog 

     

 
     

-
Version 1.3-2 (development)2024-02-06

-
  • add sqrt=TRUE in calls to Matrix determinant methods in matrix_ldet
    • 
+
    Version 1.3-3 (development)2024-05-31
    
+
    • protect errorsarlm against missing Durbin= if only intercept
      • 
+
    • fix longstanding bugs in getVmate in a non-default side logic branch
      • 
+
    • add return_impacts= to lmSLX to work around issues with aliased variables; impacts should be improved to handle this case when time permits
      • 
+
    • add sqrt=TRUE in calls to Matrix determinant methods in matrix_ldet
      • 
 
    • add igraph (>= 2.0.0) in DESCRIPTION for re-named igraph functions
      • 
 

     

diff --git a/docs/pkgdown.yml b/docs/pkgdown.yml
index 4cda807..0c5a74b 100644
--- a/docs/pkgdown.yml
+++ b/docs/pkgdown.yml
@@ -5,5 +5,5 @@ articles:
   nb_igraph: nb_igraph.html
   sids_models: sids_models.html
   SpatialFiltering: SpatialFiltering.html
-last_built: 2024-04-29T13:14Z
+last_built: 2024-05-31T15:10Z
 
diff --git a/docs/reference/ME.html b/docs/reference/ME.html
index 48ae762..838f2c3 100644
--- a/docs/reference/ME.html
+++ b/docs/reference/ME.html
@@ -168,7 +168,7 @@ 
Examples

 #> eV[,1], I: 0.08290518 ZI: NA, pr(ZI): 0.04
 #> eV[,9], I: 0.06426565 ZI: NA, pr(ZI): 0.14
 #>    user  system elapsed 
-#>   1.202   0.007   1.222 
+#>   1.512   0.005   1.527 
 glmME <- glm(c(hopkins_part) ~ 1 + fitted(MEbinom1), family="binomial")
 #anova(glmME, test="Chisq")
 coef(summary(glmME))
@@ -220,7 +220,7 @@ 
Examples

 #> eV[,6], I: 0.1178225 ZI: NA, pr(ZI): 0.08
 #> eV[,4], I: 0.06242664 ZI: NA, pr(ZI): 0.27
 #>    user  system elapsed 
-#>   0.589   0.003   0.606 
+#>   0.636   0.000   0.639 
 lagcol1
 #>   Eigenvector ZI pr(ZI)
 #> 0          NA NA   0.01
diff --git a/docs/reference/ML_models.html b/docs/reference/ML_models.html
index 4cca555..f161fa5 100644
--- a/docs/reference/ML_models.html
+++ b/docs/reference/ML_models.html
@@ -312,6 +312,9 @@ 
Control arguments

   
pre_eig

 
default NULL; may be used to pass a pre-computed vector of eigenvalues

 
+  
return_impacts

+
default TRUE; may be set FALSE to avoid problems calculating impacts with aliased variables

+
   
OrdVsign

 
default 1; used to set the sign of the final component to negative if -1 (alpha times ((sigma squared) squared) in Ord (1975) equation B.1).

 
@@ -584,7 +587,7 @@ 
Examples

 #> Computing eigenvalues ...
 #> 
 #>    user  system elapsed 
-#>   0.143   0.000   0.145 
+#>   0.214   0.000   0.216 
 summary(COL.lag.M)
 #> 
 #> Call:lagsarlm(formula = CRIME ~ INC + HOVAL, data = COL.OLD, listw = listw, 
@@ -640,7 +643,7 @@ 
Examples

 #> Computing eigenvalues ...
 #> 
 #>    user  system elapsed 
-#>   0.394   0.000   0.396 
+#>   0.564   0.003   0.571 
 summary(COL.lag.sp)
 #> 
 #> Call:lagsarlm(formula = CRIME ~ INC + HOVAL, data = COL.OLD, listw = listw, 
@@ -1421,83 +1424,83 @@ 
Examples

 #>   2.22497381 
 print(system.time(ev <- eigenw(similar.listw(listw))))
 #>    user  system elapsed 
-#>   0.001   0.000   0.001 
+#>   0.001   0.000   0.002 
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="eigen", control=list(pre_eig=ev))))
 #>    user  system elapsed 
-#>   0.143   0.000   0.143 
+#>   0.224   0.000   0.226 
 ocoef <- coefficients(COL.errW.eig)
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="eigen", control=list(pre_eig=ev, LAPACK=FALSE))))
 #>    user  system elapsed 
-#>   0.139   0.000   0.140 
+#>   0.253   0.000   0.254 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="eigen", control=list(pre_eig=ev, compiled_sse=TRUE))))
 #>    user  system elapsed 
-#>   0.151   0.000   0.153 
+#>   0.231   0.000   0.233 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix_J", control=list(super=TRUE))))
 #> Warning: the default value of argument 'sqrt' of method 'determinant(<CHMfactor>, <logical>)' may change from TRUE to FALSE as soon as the next release of Matrix; set 'sqrt' when programming
 #>    user  system elapsed 
-#>   0.159   0.000   0.160 
+#>   0.250   0.000   0.252 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix_J", control=list(super=FALSE))))
 #>    user  system elapsed 
-#>   0.158   0.000   0.160 
+#>   0.266   0.000   0.269 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix_J", control=list(super=as.logical(NA)))))
 #>    user  system elapsed 
-#>   0.158   0.000   0.159 
+#>   0.229   0.000   0.232 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix", control=list(super=TRUE))))
 #>    user  system elapsed 
-#>   0.144   0.000   0.144 
+#>   0.232   0.000   0.234 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix", control=list(super=FALSE))))
 #>    user  system elapsed 
-#>   0.144   0.000   0.144 
+#>   0.213   0.000   0.214 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="Matrix", control=list(super=as.logical(NA)))))
 #>    user  system elapsed 
-#>   0.143   0.000   0.145 
+#>   0.298   0.000   0.300 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="spam", control=list(spamPivot="MMD"))))
 #>    user  system elapsed 
-#>   0.150   0.001   0.152 
+#>   0.324   0.002   0.337 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="spam", control=list(spamPivot="RCM"))))
 #>    user  system elapsed 
-#>   0.152   0.000   0.152 
+#>   0.254   0.000   0.256 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="spam_update", control=list(spamPivot="MMD"))))
 #>    user  system elapsed 
-#>   0.149   0.000   0.149 
+#>   0.287   0.000   0.300 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 print(system.time(COL.errW.eig <- errorsarlm(CRIME ~ INC + HOVAL, data=COL.OLD,
  listw, method="spam_update", control=list(spamPivot="RCM"))))
 #>    user  system elapsed 
-#>   0.168   0.000   0.169 
+#>   0.227   0.000   0.229 
 print(all.equal(ocoef, coefficients(COL.errW.eig)))
 #> [1] TRUE
 # }
diff --git a/docs/reference/SLX.html b/docs/reference/SLX.html
index 67f8dce..3157ee8 100644
--- a/docs/reference/SLX.html
+++ b/docs/reference/SLX.html
@@ -70,7 +70,7 @@ 
Spatial Durbin linear (SLX, spatially lagged X) model

 
     

     
lmSLX(formula, data = list(), listw, na.action, weights=NULL, Durbin=TRUE,
- zero.policy=NULL)
+ zero.policy=NULL, return_impacts=TRUE)
 # S3 method for SlX
 print(x, digits = max(3L, getOption("digits") - 3L), ...)
 # S3 method for SlX
@@ -115,6 +115,9 @@ 
Arguments

   
zero.policy

 
default NULL, use global option value; if TRUE assign zero to the lagged value of zones without neighbours, if FALSE assign NA

 
+  
return_impacts

+
default TRUE; may be set FALSE to avoid problems calculating impacts with aliased variables

+
   
digits

 
the number of significant digits to use when printing

 
diff --git a/docs/reference/aple.html b/docs/reference/aple.html
index 4798cea..4a0b37e 100644
--- a/docs/reference/aple.html
+++ b/docs/reference/aple.html
@@ -149,7 +149,17 @@ 
Examples

 #> [1] 0.6601805
 # \dontrun{
 errorsarlm(yield_detrend ~ 1, wheat, spdep::nb2listw(nbr12, style="W"))
-#> Error in errorsarlm(yield_detrend ~ 1, wheat, spdep::nb2listw(nbr12, style = "W")): argument "Durbin" is missing, with no default
+#> 
+#> Call:
+#> errorsarlm(formula = yield_detrend ~ 1, data = wheat, listw = spdep::nb2listw(nbr12, 
+#>     style = "W"))
+#> Type: error 
+#> 
+#> Coefficients:
+#>      lambda (Intercept) 
+#>  0.60189687 -0.00251772 
+#> 
+#> Log likelihood: -192.9519 
 # }
 

     

diff --git a/docs/reference/invIrM.html b/docs/reference/invIrM.html
index 38ed769..fc961af 100644
--- a/docs/reference/invIrM.html
+++ b/docs/reference/invIrM.html
@@ -178,16 +178,16 @@ 
Examples

 x <- matrix(rnorm(length(nb7rt)), ncol=1)
 system.time(e <- invIrM(nb7rt, rho=0.9, method="chol", feasible=TRUE) %*% x)
 #>    user  system elapsed 
-#>   0.002   0.000   0.003 
+#>   0.003   0.000   0.003 
 system.time(e <- invIrM(nb7rt, rho=0.9, method="chol", feasible=NULL) %*% x)
 #>    user  system elapsed 
-#>   0.003   0.000   0.003 
+#>   0.002   0.000   0.003 
 system.time(e <- invIrM(nb7rt, rho=0.9, method="solve", feasible=TRUE) %*% x)
 #>    user  system elapsed 
 #>   0.001   0.000   0.001 
 system.time(e <- invIrM(nb7rt, rho=0.9, method="solve", feasible=NULL) %*% x)
 #>    user  system elapsed 
-#>   0.001   0.000   0.002 
+#>   0.002   0.000   0.001 
 # }
 

     

diff --git a/docs/reference/lagmess.html b/docs/reference/lagmess.html
index a6e501d..103ecf3 100644
--- a/docs/reference/lagmess.html
+++ b/docs/reference/lagmess.html
@@ -250,7 +250,7 @@ 
Examples

 #> 
 system.time(obj2 <- lagmess(log(PRICE) ~ PATIO + log(AGE) + log(SQFT), data=baltimore, listw=lw))
 #>    user  system elapsed 
-#>    0.03    0.00    0.03 
+#>   0.042   0.000   0.043 
 (x <- summary(obj2))
 #> Matrix exponential spatial lag model:
 #> 
diff --git a/docs/reference/sparse_mat.html b/docs/reference/sparse_mat.html
index 427107a..6e645cb 100644
--- a/docs/reference/sparse_mat.html
+++ b/docs/reference/sparse_mat.html
@@ -180,11 +180,11 @@ 
Examples

 W <- as(lw, "CsparseMatrix")
 system.time(e <- invIrM(nb7rt, rho=0.98, method="solve", feasible=NULL) %*% x)
 #>    user  system elapsed 
-#>   0.002   0.000   0.002 
+#>   0.003   0.000   0.003 
 system.time(ee <- powerWeights(W, rho=0.98, X=x))
 #> Warning: not converged within order iterations
 #>    user  system elapsed 
-#>   0.181   0.010   0.192 
+#>   0.241   0.001   0.246 
 str(attr(ee, "internal"))
 #> List of 5
 #>  $ series: num [1:250] 0.287 0.234 0.201 0.178 0.16 ...
@@ -197,10 +197,10 @@ 
Examples

 # \dontrun{
 system.time(ee <- powerWeights(W, rho=0.9, X=x))
 #>    user  system elapsed 
-#>   0.133   0.005   0.140 
+#>   0.195   0.000   0.200 
 system.time(ee <- powerWeights(W, rho=0.98, order=1000, X=x))
 #>    user  system elapsed 
-#>   0.769   0.008   0.782 
+#>   0.860   0.000   0.868 
 all.equal(e, as(ee, "matrix"), check.attributes=FALSE)
 #> [1] TRUE
 nb60rt <- spdep::cell2nb(60, 60, torus=TRUE)
@@ -209,16 +209,16 @@ 
Examples

 x <- matrix(rnorm(dim(W)[1]), ncol=1)
 system.time(ee <- powerWeights(W, rho=0.3, X=x))
 #>    user  system elapsed 
-#>   0.009   0.000   0.009 
+#>   0.011   0.000   0.011 
 str(as(ee, "matrix"))
 #>  num [1:3600, 1] -0.383 0.207 -0.731 1.552 0.32 ...
 #>  - attr(*, "dimnames")=List of 2
 #>   ..$ : chr [1:3600] "1:1" "2:1" "3:1" "4:1" ...
 #>   ..$ : NULL
 obj <- errorsarlm(as(ee, "matrix")[,1] ~ 1, listw=spdep::nb2listw(nb60rt), method="Matrix")
-#> Error in errorsarlm(as(ee, "matrix")[, 1] ~ 1, listw = spdep::nb2listw(nb60rt),     method = "Matrix"): argument "Durbin" is missing, with no default
 coefficients(obj)
-#> Error in eval(expr, envir, enclos): object 'obj' not found
+#>      lambda (Intercept) 
+#>  0.30639880  0.01380415 
 # }
 

     

diff --git a/docs/reference/spautolm.html b/docs/reference/spautolm.html
index 9f7edce..6695535 100644
--- a/docs/reference/spautolm.html
+++ b/docs/reference/spautolm.html
@@ -422,7 +422,7 @@ 
Examples

  data=nydata, listw=listw_NY, family="SAR", method="eigen",
  control=list(pre_eig=eigs)))
 #>    user  system elapsed 
-#>   0.196   0.000   0.197 
+#>   0.313   0.000   0.314 
 res <- summary(esar1f)
 print(res)
 #> 
@@ -467,7 +467,7 @@ 
Examples

 system.time(esar1M <- spautolm(Z ~ PEXPOSURE + PCTAGE65P + PCTOWNHOME,
  data=nydata, listw=listw_NY, family="SAR", method="Matrix"))
 #>    user  system elapsed 
-#>   0.221   0.000   0.221 
+#>   0.426   0.000   0.458 
 summary(esar1M)
 #> 
 #> Call: 
@@ -498,7 +498,7 @@ 
Examples

  data=nydata, listw=listw_NY, family="SAR", method="Matrix",
  control=list(super=TRUE)))
 #>    user  system elapsed 
-#>   0.196   0.000   0.196 
+#>   0.373   0.001   0.379 
 summary(esar1M)
 #> 
 #> Call: 
@@ -558,7 +558,7 @@ 
Examples

 system.time(esar1wM <- spautolm(Z ~ PEXPOSURE + PCTAGE65P + PCTOWNHOME,
  data=nydata, listw=listw_NY, weights=POP8, family="SAR", method="Matrix"))
 #>    user  system elapsed 
-#>   0.212   0.000   0.214 
+#>   0.382   0.000   0.385 
 summary(esar1wM)
 #> 
 #> Call: 
@@ -675,7 +675,7 @@ 
Examples

 system.time(ecar1M <- spautolm(Z ~ PEXPOSURE + PCTAGE65P + PCTOWNHOME,
  data=nydata, listw=listw_NY, family="CAR", method="Matrix"))
 #>    user  system elapsed 
-#>   0.236   0.000   0.237 
+#>   0.441   0.000   0.444 
 summary(ecar1M)
 #> 
 #> Call: 
@@ -737,7 +737,7 @@ 
Examples

 system.time(ecar1wM <- spautolm(Z ~ PEXPOSURE + PCTAGE65P + PCTOWNHOME,
  data=nydata, listw=listw_NY, weights=POP8, family="CAR", method="Matrix"))
 #>    user  system elapsed 
-#>   0.243   0.000   0.244 
+#>   0.437   0.001   0.456 
 summary(ecar1wM)
 #> 
 #> Call: 
diff --git a/docs/reference/trW.html b/docs/reference/trW.html
index 77af24b..89ade63 100644
--- a/docs/reference/trW.html
+++ b/docs/reference/trW.html
@@ -150,11 +150,11 @@ 
Examples

 set.seed(1100)
 system.time(trMC <- trW(W, type="MC"))
 #>    user  system elapsed 
-#>   0.005   0.000   0.006 
+#>   0.006   0.000   0.006 
 str(trMC)
 #>  num [1:30] 0 10.91 3.69 5.36 3.64 ...
 #>  - attr(*, "sd")= num [1:30] NA NA 0.598 0.495 0.489 ...
-#>  - attr(*, "timings")= Named num [1:2] 0.005 0.005
+#>  - attr(*, "timings")= Named num [1:2] 0.006 0.006
 #>   ..- attr(*, "names")= chr [1:2] "user.self" "elapsed"
 #>  - attr(*, "type")= chr "MC"
 #>  - attr(*, "n")= int 49
@@ -169,10 +169,10 @@ 
Examples

 W <- forceSymmetric(as(listwS, "CsparseMatrix"))
 system.time(trmom <- trW(listw=listwS, m=24, type="moments"))
 #>    user  system elapsed 
-#>   0.002   0.000   0.001 
+#>   0.002   0.000   0.002 
 str(trmom)
 #>  num [1:24] 0 10.91 3.65 5.62 3.66 ...
-#>  - attr(*, "timings")= Named num [1:2] 0.002 0.001
+#>  - attr(*, "timings")= Named num [1:2] 0.002 0.002
 #>   ..- attr(*, "names")= chr [1:2] "user.self" "elapsed"
 #>  - attr(*, "type")= chr "moments"
 #>  - attr(*, "n")= int 49
@@ -180,10 +180,10 @@ 
Examples

 #> [1] TRUE
 system.time(trMat <- trW(W, m=24, type="mult"))
 #>    user  system elapsed 
-#>   0.003   0.000   0.003 
+#>   0.005   0.000   0.005 
 str(trMat)
 #>  num [1:24] 0 10.91 3.65 5.62 3.66 ...
-#>  - attr(*, "timings")= Named num [1:2] 0.003 0.003
+#>  - attr(*, "timings")= Named num [1:2] 0.005 0.005
 #>   ..- attr(*, "names")= chr [1:2] "user.self" "elapsed"
 #>  - attr(*, "type")= chr "mult"
 #>  - attr(*, "n")= int 49
@@ -192,11 +192,11 @@ 
Examples

 set.seed(1)
 system.time(trMC <- trW(W, m=24, type="MC"))
 #>    user  system elapsed 
-#>   0.006   0.000   0.005 
+#>   0.009   0.000   0.009 
 str(trMC)
 #>  num [1:24] 0 10.91 2.44 4.97 2.82 ...
 #>  - attr(*, "sd")= num [1:24] NA NA 0.618 0.501 0.451 ...
-#>  - attr(*, "timings")= Named num [1:2] 0.006 0.005
+#>  - attr(*, "timings")= Named num [1:2] 0.008 0.009
 #>   ..- attr(*, "names")= chr [1:2] "user.self" "elapsed"
 #>  - attr(*, "type")= chr "MC"
 #>  - attr(*, "n")= int 49
@@ -206,10 +206,10 @@ 
Examples

 listwS <- similar.listw(listw)
 system.time(trmom <- trW(listw=listwS, m=24, type="moments"))
 #>    user  system elapsed 
-#>   0.108   0.000   0.108 
+#>   0.128   0.000   0.128 
 str(trmom)
 #>  num [1:24] 0 124.2 32.7 63.7 33.2 ...
-#>  - attr(*, "timings")= Named num [1:2] 0.108 0.108
+#>  - attr(*, "timings")= Named num [1:2] 0.128 0.128
 #>   ..- attr(*, "names")= chr [1:2] "user.self" "elapsed"
 #>  - attr(*, "type")= chr "moments"
 #>  - attr(*, "n")= int 506
@@ -225,7 +225,7 @@ 
Examples

 }
 system.time(trmomp <- trW(listw=listwS, m=24, type="moments"))
 #>    user  system elapsed 
-#>   0.111   0.000   0.112 
+#>   0.125   0.000   0.127 
 if(!get.mcOption()) {
   set.ClusterOption(NULL)
   stopCluster(cl)