ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

Cluto presentation

?Download as PPTX, PDF?
?
Roseline Antai
Roseline Antai

CLUTO is a software toolkit used for clustering high-dimensional datasets and analyzing cluster characteristics. It contains two main algorithms: Vcluster, which clusters based on the actual multi-dimensional data representation, and Scluster, which clusters based on a pre-computed similarity matrix. CLUTO can be run from the command line with various optional parameters to control the clustering method, analysis, and visualization of results.

1 of 16
Downloaded 22 times
CLUTO A Clustering Toolkit By Roseline Antai
? CLUTO is a software package which is used for clustering high dimensional datasets and for analyzing the characteristics of the various clusters.
Algorithms of CLUTO ? Vcluster ? Scluster Major difference: Input Vcluster: actual mutidimensional representation of the objects to be clustered. Scluster: The similarity matrix (or graph) between these objects.
Calling Sequence vcluster [optional parameters] MatrixFile Nclusters scluster [optional parameters] MatrixFile NClusters
Optional Parameters ? Standard specification -paramname or ¨Cparamname = value ? Three categories: ¨C Clustering algorithm parameters ¨C Reporting and Analysis parameters ¨C Cluster Visualization parameters
Clustering algorithm parameters ? Control how CLUTO computes the clustering solution. ? Examples 1. -clmethod=string ( rb, agglo,direct,graph, etc) 2. -sim = string (cos,corr,dist,jacc) 3. -crfun = string (i1,i2 etc) 4. -fulltree
Reporting and Analysis Parameters ? Control the amount of information that vcluster and scluster report about the clusters as well as the analysis performed on discovered clusters. ? Examples 1. -clustfile = string. ( Default is MatrixFile.clustering.Nclusters( or GraphFile)) 2. -clabelfile = string (name of the file that¡¯s stores the labels of the columns. Used when ¨Cshowfeatues, - showsummaries or ¨Clabeltree are used)
3. -rlabelfile=string 4. -rclassfile=string (Stores the labels of the rows ¨C objects to be clustered). 5. -showtree 6. -showfeatures (descriptive and discriminating)
Cluster Visualization Parameters ? Simple plots of the original input matrix which show how the different objects (rows) and features (columns) are clustered together. ? Examples 1. -plottree = string; gives graphic representation of the entire hierarchical tree 2. -plotmatrix = string; shows how the rows of the original matrix are clustered together.
A practical example ¨C ../cluto/Linux/vcluster -clmethod=rb -sim=cos - fulltree -rlabelfile=Final_Results/rlabelfile - rclassfile=Final_Results/classfile -showtree - plotformat=gif -plottree=Final_Results/Images/PT- Final10d -plotmatrix=Final_Results/Images/PM- Final10d -plotclusters=Final_Results/Images/PC- Final10d -showfeatures Final_Results/FinalOutput10d-Vt.mat 4
roselineantai@ubuntu:~/JLSI/jlsi$ ./clusterscript2.sh vcluster (CLUTO 2.1.1) Copyright 2001-03, Regents of the University of Minnesota Matrix Information ----------------------------------------------------------- Name: Final_Results2/FinalOutput50dFinal.mat, #Rows: 59, #Columns: 59, #NonZeros: 3481 Options ---------------------------------------------------------------------- CLMethod=RB, CRfun=I2, SimFun=Cosine, #Clusters: 4 RowModel=None, ColModel=None, GrModel=SY-DIR, NNbrs=40 Colprune=1.00, EdgePrune=-1.00, VtxPrune=-1.00, MinComponent=5 CSType=Best, AggloFrom=0, AggloCRFun=I2, NTrials=10, NIter=10 Solution --------------------------------------------------------------------- ------------------------------------------------------------------------ 4-way clustering: [I2=5.65e+01] [59 of 59], Entropy: 0.438, Purity: 0.729 ------------------------------------------------------------------------ cid Size ISim ISdev ESim ESdev Entpy Purty | Sem Imp Deo Evo ------------------------------------------------------------------------ 0 2 +0.779 +0.000 +0.638 +0.222 0.000 1.000 | 0 2 0 0 1 15 +0.918 +0.024 +0.853 +0.021 0.453 0.800 | 1 0 2 12 2 22 +0.923 +0.015 +0.866 +0.019 0.416 0.818 | 1 18 3 0 3 20 +0.923 +0.015 +0.869 +0.034 0.496 0.550 | 9 0 11 0 ------------------------------------------------------------------------ -------------------------------------------------------------------------------- 4-way clustering solution - Descriptive & Discriminating Features... -------------------------------------------------------------------------------- Cluster 0, Size: 2, ISim: 0.779, ESim: 0.638 Descriptive: col00010 32.4%, col00039 22.7%, col00013 5.6%, col00019 3.9%, col00047 2.2% Discriminating: col00010 53.5%, col00039 22.3%, col00013 2.5%, col00020 2.1%, col00012 1.4% Cluster 1, Size: 15, ISim: 0.918, ESim: 0.853 Descriptive: col00013 11.3%, col00006 8.2%, col00019 7.9%, col00058 4.4%, col00047 4.2% Discriminating: col00006 14.8%, col00007 7.4%, col00031 7.1%, col00029 6.6%, col00053 6.2% Cluster 2, Size: 22, ISim: 0.923, ESim: 0.866 Descriptive: col00013 14.3%, col00020 7.1%, col00014 6.4%, col00019 4.4%, col00047 4.4% Discriminating: col00020 7.0%, col00014 5.9%, col00013 5.5%, col00006 5.1%, col00010 4.3% Cluster 3, Size: 20, ISim: 0.923, ESim: 0.869 Descriptive: col00013 8.9%, col00047 6.6%, col00019 5.2%, col00020 4.4%, col00014 3.8% Discriminating: col00015 8.0%, col00013 7.4%, col00046 5.2%, col00006 5.0%, col00042 4.3% -------------------------------------------------------------------------------- ------------------------------------------------------------------------------ Hierarchical Tree that optimizes the I2 criterion function... ------------------------------------------------------------------------------ Sem Imp Deo Evo ------------------------------------ 6 |---4 | |---2 1 18 3 0 | |---3 9 0 11 0 |-5 |-----1 1 0 2 12 |-----0 0 2 0 0 ------------------------------------ ------------------------------------------------------------------------------ Timing Information ----------------------------------------------------------- I/O: 0.004 sec Clustering: 0.008 sec Reporting: 0.316 sec
Classfile and rlabelfile 0 Evo 1 Sem Imp 2 Imp 3 Deo Deo 4 Imp 5 Imp Deo 6 Deo 7 Imp Deo 8 Deo 9 Imp Sem 10 Deo 11 Sem Imp 12 Imp 13 Evo 14 15
Cluto presentation
Cluto presentation
The plot uses red to denote positive values and green to denote negative values. Bright red/green indicate large positive/negative values, whereas colors close to white indicate values close to zero.
Sem 0 Imp 1 Deo 2 Deo 3 Sem 4 Evo 5 Evo 6 Sem 7 Imp 8 Imp 9 Deo 10 Deo 11 Imp 12 Imp 13 Deo 14 Deo 15 Imp 16 Deo 17 Deo 18 Imp 19 Sem 20

More Related Content

Cluto presentation

  • 1. CLUTO A Clustering Toolkit By Roseline Antai
  • 2. ? CLUTO is a software package which is used for clustering high dimensional datasets and for analyzing the characteristics of the various clusters.
  • 3. Algorithms of CLUTO ? Vcluster ? Scluster Major difference: Input Vcluster: actual mutidimensional representation of the objects to be clustered. Scluster: The similarity matrix (or graph) between these objects.
  • 4. Calling Sequence vcluster [optional parameters] MatrixFile Nclusters scluster [optional parameters] MatrixFile NClusters
  • 5. Optional Parameters ? Standard specification -paramname or ¨Cparamname = value ? Three categories: ¨C Clustering algorithm parameters ¨C Reporting and Analysis parameters ¨C Cluster Visualization parameters
  • 6. Clustering algorithm parameters ? Control how CLUTO computes the clustering solution. ? Examples 1. -clmethod=string ( rb, agglo,direct,graph, etc) 2. -sim = string (cos,corr,dist,jacc) 3. -crfun = string (i1,i2 etc) 4. -fulltree
  • 7. Reporting and Analysis Parameters ? Control the amount of information that vcluster and scluster report about the clusters as well as the analysis performed on discovered clusters. ? Examples 1. -clustfile = string. ( Default is MatrixFile.clustering.Nclusters( or GraphFile)) 2. -clabelfile = string (name of the file that¡¯s stores the labels of the columns. Used when ¨Cshowfeatues, - showsummaries or ¨Clabeltree are used)
  • 8. 3. -rlabelfile=string 4. -rclassfile=string (Stores the labels of the rows ¨C objects to be clustered). 5. -showtree 6. -showfeatures (descriptive and discriminating)
  • 9. Cluster Visualization Parameters ? Simple plots of the original input matrix which show how the different objects (rows) and features (columns) are clustered together. ? Examples 1. -plottree = string; gives graphic representation of the entire hierarchical tree 2. -plotmatrix = string; shows how the rows of the original matrix are clustered together.
  • 10. A practical example ¨C ../cluto/Linux/vcluster -clmethod=rb -sim=cos - fulltree -rlabelfile=Final_Results/rlabelfile - rclassfile=Final_Results/classfile -showtree - plotformat=gif -plottree=Final_Results/Images/PT- Final10d -plotmatrix=Final_Results/Images/PM- Final10d -plotclusters=Final_Results/Images/PC- Final10d -showfeatures Final_Results/FinalOutput10d-Vt.mat 4
  • 11. roselineantai@ubuntu:~/JLSI/jlsi$ ./clusterscript2.sh vcluster (CLUTO 2.1.1) Copyright 2001-03, Regents of the University of Minnesota Matrix Information ----------------------------------------------------------- Name: Final_Results2/FinalOutput50dFinal.mat, #Rows: 59, #Columns: 59, #NonZeros: 3481 Options ---------------------------------------------------------------------- CLMethod=RB, CRfun=I2, SimFun=Cosine, #Clusters: 4 RowModel=None, ColModel=None, GrModel=SY-DIR, NNbrs=40 Colprune=1.00, EdgePrune=-1.00, VtxPrune=-1.00, MinComponent=5 CSType=Best, AggloFrom=0, AggloCRFun=I2, NTrials=10, NIter=10 Solution --------------------------------------------------------------------- ------------------------------------------------------------------------ 4-way clustering: [I2=5.65e+01] [59 of 59], Entropy: 0.438, Purity: 0.729 ------------------------------------------------------------------------ cid Size ISim ISdev ESim ESdev Entpy Purty | Sem Imp Deo Evo ------------------------------------------------------------------------ 0 2 +0.779 +0.000 +0.638 +0.222 0.000 1.000 | 0 2 0 0 1 15 +0.918 +0.024 +0.853 +0.021 0.453 0.800 | 1 0 2 12 2 22 +0.923 +0.015 +0.866 +0.019 0.416 0.818 | 1 18 3 0 3 20 +0.923 +0.015 +0.869 +0.034 0.496 0.550 | 9 0 11 0 ------------------------------------------------------------------------ -------------------------------------------------------------------------------- 4-way clustering solution - Descriptive & Discriminating Features... -------------------------------------------------------------------------------- Cluster 0, Size: 2, ISim: 0.779, ESim: 0.638 Descriptive: col00010 32.4%, col00039 22.7%, col00013 5.6%, col00019 3.9%, col00047 2.2% Discriminating: col00010 53.5%, col00039 22.3%, col00013 2.5%, col00020 2.1%, col00012 1.4% Cluster 1, Size: 15, ISim: 0.918, ESim: 0.853 Descriptive: col00013 11.3%, col00006 8.2%, col00019 7.9%, col00058 4.4%, col00047 4.2% Discriminating: col00006 14.8%, col00007 7.4%, col00031 7.1%, col00029 6.6%, col00053 6.2% Cluster 2, Size: 22, ISim: 0.923, ESim: 0.866 Descriptive: col00013 14.3%, col00020 7.1%, col00014 6.4%, col00019 4.4%, col00047 4.4% Discriminating: col00020 7.0%, col00014 5.9%, col00013 5.5%, col00006 5.1%, col00010 4.3% Cluster 3, Size: 20, ISim: 0.923, ESim: 0.869 Descriptive: col00013 8.9%, col00047 6.6%, col00019 5.2%, col00020 4.4%, col00014 3.8% Discriminating: col00015 8.0%, col00013 7.4%, col00046 5.2%, col00006 5.0%, col00042 4.3% -------------------------------------------------------------------------------- ------------------------------------------------------------------------------ Hierarchical Tree that optimizes the I2 criterion function... ------------------------------------------------------------------------------ Sem Imp Deo Evo ------------------------------------ 6 |---4 | |---2 1 18 3 0 | |---3 9 0 11 0 |-5 |-----1 1 0 2 12 |-----0 0 2 0 0 ------------------------------------ ------------------------------------------------------------------------------ Timing Information ----------------------------------------------------------- I/O: 0.004 sec Clustering: 0.008 sec Reporting: 0.316 sec
  • 12. Classfile and rlabelfile 0 Evo 1 Sem Imp 2 Imp 3 Deo Deo 4 Imp 5 Imp Deo 6 Deo 7 Imp Deo 8 Deo 9 Imp Sem 10 Deo 11 Sem Imp 12 Imp 13 Evo 14 15
  • 15. The plot uses red to denote positive values and green to denote negative values. Bright red/green indicate large positive/negative values, whereas colors close to white indicate values close to zero.
  • 16. Sem 0 Imp 1 Deo 2 Deo 3 Sem 4 Evo 5 Evo 6 Sem 7 Imp 8 Imp 9 Deo 10 Deo 11 Imp 12 Imp 13 Deo 14 Deo 15 Imp 16 Deo 17 Deo 18 Imp 19 Sem 20