diff --git a/404.html b/404.html index d36b44c..3f516c1 100644 --- a/404.html +++ b/404.html @@ -32,7 +32,7 @@
diff --git a/LICENSE-text.html b/LICENSE-text.html index 4115c44..ca27cb9 100644 --- a/LICENSE-text.html +++ b/LICENSE-text.html @@ -17,7 +17,7 @@ diff --git a/articles/CytoMDS.html b/articles/CytoMDS.html index 6f56a53..138e671 100644 --- a/articles/CytoMDS.html +++ b/articles/CytoMDS.html @@ -33,7 +33,7 @@ @@ -153,7 +153,7 @@Hauchamps P, Gatto L (2023). CytoMDS: Low Dimensions projection of cytometry samples. -R package version 0.99.2, https://uclouvain-cbio.github.io/CytoMDS. +R package version 0.99.3, https://uclouvain-cbio.github.io/CytoMDS.
@Manual{, title = {CytoMDS: Low Dimensions projection of cytometry samples}, author = {Philippe Hauchamps and Laurent Gatto}, year = {2023}, - note = {R package version 0.99.2}, + note = {R package version 0.99.3}, url = {https://uclouvain-cbio.github.io/CytoMDS}, }diff --git a/index.html b/index.html index efd0e0f..ffbd5f0 100644 --- a/index.html +++ b/index.html @@ -33,7 +33,7 @@ diff --git a/news/index.html b/news/index.html index ac9ed16..dc5738f 100644 --- a/news/index.html +++ b/news/index.html @@ -17,7 +17,7 @@ @@ -63,6 +63,11 @@
cmdscale
alike, aka Torgerson's algorithm),
but is the SMACOF algorithm for metric distances that are not
necessarily euclidean.
-Note that after the obtention of the projections on the nDim
dimensions,
+After having obtained the projections on the nDim
dimensions,
we always apply svd decomposition to visualize as first axes the ones that
-contain the most variance of the projected dataset in nDim
dimensions
+contain the most variance of the projected dataset in nDim
dimensions.
+Instead of being provided directly by the user, the nDim
parameter can
+otherwise be found iteratively by finding the minimum nDim
parameter that
+allows the projection to reach a target pseudo RSquare.
+If this is the case, the maxDim
parameter is used to avoid
+looking for too big projection spaces.
computeMetricMDS(pwDist, nDim = 2, seed = NULL, ...)
computeMetricMDS(
+ pwDist,
+ nDim = NULL,
+ seed = NULL,
+ targetPseudoRSq = 0.99,
+ maxDim = 128,
+ ...
+)
number of dimensions of projection, as input to SMACOF algorithm
number of dimensions of projection, as input to SMACOF algorithm
+if not provided, will be found iteratively using targetPseudoRSq
init=="random"
but not only)
+target pseudo RSquare to be reached
+(only used when nDim
is set to NULL)
in case nDim
is found iteratively,
+maximum number of dimensions the search procedure is allowed to explore
additional parameters passed to SMACOF algorithm