----------------------------------------------------------------------CLUSTERBASICS   README    Alexander Strehl    Version 1.0   2002-04-20----------------------------------------------------------------------This package contains MATLAB implementation of a variety of clusterfunctions as described in       A. Strehl, J. Ghosh and R. Mooney, "Impact of Similarity      Measures on Web-page Clustering", Proc. of the 17th National      Conference on Artificial Intelligence: Workshop of Artificial      Intelligence for Web Search (AAAI 2000), July 2000, Austin,      Texas, pp. 58-64 and      A. Strehl and J. Ghosh, "Relationship-based Clustering and      Visualization for High-dimensional Data Mining", Special      Issue on Mining Web-based Data for E-Business Applications      of the INFORMS Journal on Computing, 2002----------------------------------------------------------------------To use it copy all 31 files      ClusterBasicsTest.m      README_ClusterBasics.txt      checkcl.m      checks.m      clagglmin.m      clcgraph.m      clgraph.m      clhgraph.m      clkmeans.m      clrand.m      clucent.m      cmetis.m      evalbalance.m      evalf.m      evalmse.m      evalmutual.m      fastchangem.m      hmetis.m      ints.m      kmeans.m      mapasc.m      mapdense.m      metis.m      onetomax.m      sgraph.m      simbjac.m      simcorr.m      simcosi.m      simeucl.m      simxjac.m      wgraph.mto a directory in your MATLAB path or the current directory inyour MATLAB environment. Also, make sure that      pmetis      shmetisor      pmetis.exe      shmetis.exeare in your machine's path and executable. Metis is a separateprogram available from      http://www-users.cs.umn/edu/~karypis/metis/download.htmlThe Linux & Win32 binaries are included with this distribution. Then,in MATLAB just type      ClusterBasicsTest;to run a test of all the functions. Type      help ClusterBasicsTest to obain other info or      edit ClusterBasicsTestto see the function calls performed for the test.All functions have been tested on LNX86 5.2.0.3084.----------------------------------------------------------------------The following functions are intended for user invocation. They arearranged into four groups of activities:* Similarity Matrix Computation. The following functions computea matrix containing all pairwise similarities between the rowvectors in a and b:      simeucl(a,b)      simcosi(a,b)      simcorr(a,b)      simbjac(a,b)      simxjac(a,b)* Clustering Algorithms. The following functions compute thecluster labels from 1 to k for all the objects described by rowvectors in x using the similarity semantic sfct (e.g., 'simeucl'):      clagglmin(x,k,sfct)      clcgraph(x,k,sfct)      clgraph(x,k,sfct)      clhgraph(x,k,sfct)      clkmeans(x,k,sfct,oldcl)      clrand(x,k,sfct)* Cluster Quality Evaluation Functions. The following functionsevaluate the quality of a clustering cl given a human imposedcategorization trueclass for the data matrix x using similaritysemantic sfct:      evalbalance(trueclass,cl,x,sfct)      evalf(trueclass,cl,x,sfct)      evalmse(trueclass,cl,x,sfct)      evalmutual(trueclass,cl,x,sfct)* Data Cleaning. The following functions can be used tocheck and correct problems with a clustering cl or a similaritymatrix s:      checkcl(cl)      checks(s)Please note that not all functions use all parameters.  Please type'help functionname' for more information on an individual function.----------------------------------------------------------------------For questions, comments & services please contact      alexander@strehl.comThis package and potential updates are available from      http://strehl.com----------------------------------------------------------------------Finally, copyright (c) 1998-2002 by Alexander Strehl. All rightsreserved.  License is granted to copy, to use, and to make and to usederivative works for research purposes, provided that the AlexanderStrehl copyright notice and this license notice is included in allcopies and any derivatives works and in all relateddocumentation. Alexander Strehl grants no other licenses expressed orimplied and the licensee acknowleges that Alexander Strehl has noliability for licensee's use or for any derivative works made bylicensee. This software is provided as is. Alexander Strehl disclaimsand licensee agrees that all warranties, express or implied, includingwithout limitation the implied warranties of merchantability andfitness for a particular purpose. Notwithstanding any other provisioncontained herein, any liability for damages resulting from thesoftware or its use is expressly disclaimed, including consequentialor any other indirect damages, whether arising in contract, tort(including negligence) or strict liability, even if Alexander Strehlis advised of the possibility of such damages. Enjoy!----------------------------------------------------------------------