X-Git-Url: https://www.fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=clueless-kmeans.git;a=blobdiff_plain;f=README.md;fp=README.md;h=53fa4ea70ff48e2e5457ec67caf91be2f5373ff7;hp=0000000000000000000000000000000000000000;hb=84a44ea96568693cce7689ffe7b65bfa23341c79;hpb=335ae9bf91a7ba82163136c88c80af8a8da6a5e7

diff --git a/README.md b/README.md
new file mode 100644
index 0000000..53fa4ea
--- /dev/null
+++ b/README.md
@@ -0,0 +1,25 @@
+# Introduction
+
+This procedure is a variant of k-means using labelled samples, which
+enforces in every cluster the same proportion of samples from every
+class. This ensures that the resulting clusters are totally
+non-informative about the class, while maximally informative about
+the signal.
+
+You can get a
+[`short report on the method,`](https://fleuret.org/papers/fleuret-clueless-kmeans2015.pdf).
+
+# Installation
+
+Executing
+
+```
+./test.sh
+```
+
+will compile the source, run the algorithm on a 2d toy example, and
+produce three graphs
+([`result-standard.png,`](https://fleuret.org/git-extract/clueless-kmeans/result-standard.png)
+[`result-clueless.png,`](https://fleuret.org/git-extract/clueless-kmeans/result-clueless.png)
+and [`result-clueless-absolute.png`](https://fleuret.org/git-extract/clueless-kmeans/result-clueless-absolute.png)) if you
+have [`gnuplot`](http://www.gnuplot.info/) installed.