version 0.1-8 (Feb.2011):
  first public release and distribution via WWW

version 0.1-9 (2011-03-31)
 - included a function to compute the adjusted Rand index of two partitions
 - included the density-based Silhouette plot as possible option when 
 displaying objects of class pdfCluster

version 0.1-10 (2011-06-10)
 - fixed a bug in pdfClassification
 - included the possibility to match argument in choosing a method for plotting
 objects of class kepdf
 - changed the default colors to be used when plotting objects of class 
 pdfSilhouette 
 - added the Call of the function in the Show-methods for function pdfCluster
 - fixed a bug in function dbs, occurring when some posterior probabilities 
 are numerically estimated as zero 

version 0.1-11 (2011-06-16)
 - fixed a bug in the procedure for finding high-density connected regions, 
 occurring when groups are so well separated that more than one connected 
 region are founded for any of the n.grid thresholds of the density estimate
 - included a message to be printed when classification of low-density data 
 points is accomplished in less than n.stage stages. 

version 0.1-12 (2011-06-27)
 - found a bug in function "quantile". Given a vector of ordered probabilities,
 it may return unsorted quantiles (at the 16th digit). Fixed by rounding to the
 15th digit the values of function "quantile".
 - some minor changes in documentation. 

version 0.1-13 (2011-06-29)
 - extended the class "dendrogram" to the slot "tree" of objects of class 
 pdfCluster, also when 1 cluster only is found. As a consequence,
 plotting the cluster tree is now possible even in this case
 - removed the slot "ctrl" from objects of class pdfCluster
 - included the slot "noc" in objects of class pdfCluster
 - included a control in pdfClassification: outliers with zero estimated 
 conditional density of belonging to the detected groups are now excluded from
 the classification  

version 1.0-0 (2012-04-20)
  Warning: major modifications of the pdfCluster package. 
  Old codes might not work or produce different results!
 - new option which implements adaptive kernel density estimation; 
 - new function hprop2f to automatically select the matrix of badwidths
 according to Silverman's(1986) approach;
 - new arguments "bwtype", "hx", "alpha" in function kepdf;
 - new kernel function "t7" to save time in computation for very 
 large datasets;
 - new arghument "kernel" in function "kepdf";
 - removed argument "h.weigths" in function "kepdf";
 - new slots "kernel", "bwtype" and "par" in objects of class "kepdf";
 - removed slots "h" and "h.weigths" in objects of class "kepdf";
 - new argument "indcol" in function "plot.kepdfdd";
 - new procedure to find connected components; to  be used instead
 of the Delaunay triangulation for high-dimensional data;
 - new arguments in function "pdfCluster": "graphtype", "Q", "lambda",
 "grid.pairs";
 - arguments in function pdfCluster left as optional only: "h", "n.stage"; 
 - new slots in objects of class pdfCluster: "pdf", "graph";
 - new arguments "cores" and "se" in function "pdfClassification";
 - new "pdfCluster" methods for objects of class "numeric", "matrix",
 "data.frame", "pdfCluster";
 - removed argument "h.funct" in pdfClassification;
 - classes "pdf.Silhouette" and "pdf.Silhouette.summary" renamed as "dbs" and
 "dbs.summary" respectively;
 - renamed slot "nc" to "noc" in dbs-class objects

version 1.0-1 (2013-07-28)
 - new default value for argument "n.grid" of function "pdfCluster", set to an 
	increasing function of the sample size 
 - added a warning message when "pdfCluster" is run with data of dimension > 6 
	and graphtype="delaunay"
 - added a warning message when "pdfCluster" is run with discrete data
 - some typos corrected in the help files and in "show" method of objects of class 
   "kepdf"
 - changed message to be displayed when "n.grid" is set to a too small value
 - added specific argument "hmult" of function "pdfCluster" (previously optional)
 - signatures of "pdfCluster" methods modified 
 - included option to break running of foreign .C function to find pairwise connections
 - added new function "groups" to extract groups; slot "clusters" in objects 
   of pdfCluster-class is now to be considered obsolete

version 1.0-2 (2014-04-28)
 - added citation to article on the package published on Journal of Statistical Software
 - removed obsolete argument DUP in call to foreign functions

version 1.0-3 (2018-12-04)
 - S3 methods registered
 - removed unworking links
 - removed warning message when the package is uploaded
 
version 1.0-4 (2022-12-01)
 - removed some warnings in package check