https://packaging.python.org/guides/supporting-windows-using-appveyor/

https://docs.anaconda.com/anaconda/packages/pkg-docs/:

To request a package not listed on this page, please create an issue on the Anaconda issues page.

and submit to CRAN, make code reusable between R and Python

Add support for the "smoothed" single linkage (OWAs over few nearest neighbours) as described in our manuscript submitted for publication in Information Sciences.

• https://cran.r-project.org/web/views/Robust.html

• https://cran.r-project.org/web/views/Cluster.html

• add to README

• add to sphinx/r

seek first pair of sets that yields g <= threshold, otherwise, choose arg min g

Also make sure dentrograms are plotted correctly in the presence of noise points

Also add examples in the tutorials

from https://github.com/gagolews/genie/blob/master/src/hclust2_distance.cpp

https://docs.travis-ci.com/user/languages/python/

Add genieclust.plots.abline() -- line R's abline()

Computing mst_dist, mst_ind is the slowest part

Getting nn_dist, nn_ind (needed to compute d_core before mst_dist, mst_ind) is also slow (M>1)

compute_full_tree, postprocess, gini_threshold, n_clusters are only applied after mst_dist, mst_ind are computed

Changing X, M, affinity, cast_float32, exact require mst_dist, mst_ind be recomputed

However, decreasing M only does not invalidate nn_dist, nn_ind - at least d_core could be generated faster.

rewrite

rewrite

https://github.com/gagolews/genie/blob/master/src/hclust2_result.h
https://github.com/gagolews/genie/blob/master/src/hclust2_result.cpp

merge matrix

?hclust Value in R

 merge: an n-1 by 2 matrix.  Row i of ‘merge’ describes the merging
          of clusters at step i of the clustering.  If an element j in
          the row is negative, then observation -j was merged at this
          stage.  If j is positive then the merge was with the cluster
          formed at the (earlier) stage j of the algorithm.  Thus
          negative entries in ‘merge’ indicate agglomerations of
          singletons, and positive entries indicate agglomerations of
          non-singletons.

  height: a set of n-1 real values (non-decreasing for ultrametric
          trees).  The clustering _height_: that is, the value of the
          criterion associated with the clustering ‘method’ for the
          particular agglomeration.

   order: a vector giving the permutation of the original observations
          suitable for plotting, in the sense that a cluster plot using
          this ordering and matrix ‘merge’ will not have crossings of
          the branches.

e.g.,
http://www.cs.ubc.ca/research/flann/
https://github.com/facebookresearch/faiss or
https://github.com/spotify/annoy
https://github.com/nmslib/nmslib [NN-descent...]
https://github.com/nmslib/hnswlib ; some will also enable
sparse input data (e.g., for text mining applications)

@todo: make the code portable

add an option to standardise/robust standardise data?

Hi,

this is a very nice package for performing hierarchical clustering. I see that you have licenced is as GPL. At the moment a large part of the python ML community (see the scikit-* family) releases BSD-licenced libraries to allow commercial usage of code, which helps build a business ecosystem around a library. Would you consider licencing your library as BSD?

See paper by M.Rezaei and P.Franti

rewrite

for example Ic values at each step

implement the
__setstate__(self, state) and
__getstate__(self) methods.