| 1 |
\import{mcx.zmm} |
| 2 |
|
| 3 |
\begin{pud::man}{ |
| 4 |
|
| 5 |
{name}{clm mate} |
| 6 |
{html_title}{The clm mate manual} |
| 7 |
{author}{Stijn van Dongen} |
| 8 |
{section}{1} |
| 9 |
{synstyle}{long} |
| 10 |
{defstyle}{long} |
| 11 |
|
| 12 |
\man_share |
| 13 |
} |
| 14 |
|
| 15 |
\${html}{\"pud::man::maketoc"} |
| 16 |
|
| 17 |
\sec{name}{NAME} |
| 18 |
\NAME{clm_mate}{compute best matches between two clusterings} |
| 19 |
|
| 20 |
\disclaim_clm{mate} |
| 21 |
|
| 22 |
\sec{synopsis}{SYNOPSIS} |
| 23 |
\par{ |
| 24 |
\clm{mate} |
| 25 |
\synoptopt{-o}{fname}{output file name} |
| 26 |
\synoptopt{-b}{omit headers} |
| 27 |
\synoptopt{--one-to-many}{require multiple hits in <clfile1>} |
| 28 |
\stdsynopt |
| 29 |
<clfile1> <clfile2> |
| 30 |
} |
| 31 |
|
| 32 |
\sec{description}{DESCRIPTION} |
| 33 |
|
| 34 |
\par{ |
| 35 |
\clm{mate} computes for each cluster \v{X} in \v{clfile1} all clusters |
| 36 |
\v{Y} in \v{clfile2} that have non-empty intersection and outputs |
| 37 |
a line with the data points listed below.} |
| 38 |
|
| 39 |
\verbatim{\:/ |
| 40 |
overlap(X,Y) # 2 * size(meet(X,Y)) / (size(X)+size(Y)) |
| 41 |
index(X) # name of cluster |
| 42 |
index(Y) # name of cluster |
| 43 |
size(meet(X,Y)) |
| 44 |
size(X-Y) # size of left difference |
| 45 |
size(Y-X) # size of right difference |
| 46 |
size(X) |
| 47 |
size(Y) |
| 48 |
projection(X, clfile2) # see below |
| 49 |
projection(Y, clfile1) # see below |
| 50 |
} |
| 51 |
|
| 52 |
\par{ |
| 53 |
The projected size of a cluster \v{X} relative to a clustering \v{K} is |
| 54 |
simply the sum of all the nodes shared between any cluster \v{Y} in \v{K} |
| 55 |
and \v{X}, duplications allowed. For example, the projected size of |
| 56 |
\v{(0,1)} relative to \v{\{(0,2,4), (1,4,9), (1,3,5)\}} equals \v{3}.} |
| 57 |
|
| 58 |
\par{ |
| 59 |
The overlap between \v{X} and \v{Y} is exactly |
| 60 |
1.0 if the two clusters are identical, and for nearly identical |
| 61 |
clusterings the score will be close to 1.0.} |
| 62 |
|
| 63 |
\par{ |
| 64 |
All of this information can also be obtained from the |
| 65 |
contingency matrix defined for two clusterings. |
| 66 |
The \v{[i,j]} row-column entry in a contigency matrix between |
| 67 |
to clusterings gives the number of entries in the intersection |
| 68 |
between cluster\~\v{i} and cluster\~\v{j} from the respective |
| 69 |
clusterings. The other information is implicitly present; |
| 70 |
the total number of nodes in clusters\~\v{i} and\~\v{j} |
| 71 |
for example can be obtained as the sum of entries in row\~\v{i} |
| 72 |
and column\~\v{j} respectively, and the difference counts |
| 73 |
can then be obtained by substracting the intersection count. |
| 74 |
The contingency matrix can easily be computed using \mcx; |
| 75 |
e.g.} |
| 76 |
|
| 77 |
\verbatim{ |
| 78 |
mcx /clfile2 lm /clfile1 lm tp mul /ting wm} |
| 79 |
|
| 80 |
\car{ |
| 81 |
will create the contingency matrix in mcl matrix format |
| 82 |
in the file \v{ting}, where columns range over the clusters |
| 83 |
in \v{clfile1}.} |
| 84 |
|
| 85 |
\par{ |
| 86 |
The output can be put to good use by sorting it numerically on |
| 87 |
that first score field. It is advisable to use a stable sort routine |
| 88 |
(use the \usearg{-s} option for UNIX sort) |
| 89 |
From this information one can quickly extract the closest |
| 90 |
clusters between two clusterings.} |
| 91 |
|
| 92 |
\sec{options}{OPTIONS} |
| 93 |
|
| 94 |
\'begin{itemize}{\mcx_itemopts} |
| 95 |
|
| 96 |
\item{\defopt{-o}{fname}{output file name}} |
| 97 |
\car{Specify the name of the output file.} |
| 98 |
\item{\defopt{-b}{omit headers}} |
| 99 |
\car{Batch mode, omit column names.} |
| 100 |
\item{\defopt{--one-to-many}{require multiple hits in <clfile1>}} |
| 101 |
\car{Do not output information for clusters in the first file |
| 102 |
that are subset of a cluster in the second file.} |
| 103 |
\stddefopt |
| 104 |
|
| 105 |
\'end{itemize} |
| 106 |
|
| 107 |
\sec{author}{AUTHOR} |
| 108 |
\par{ |
| 109 |
Stijn van Dongen. |
| 110 |
} |
| 111 |
|
| 112 |
|
| 113 |
\sec{seealso}{SEE ALSO} |
| 114 |
\par{ |
| 115 |
\mysib{mclfamily} for an overview of all the documentation |
| 116 |
and the utilities in the mcl family. |
| 117 |
} |
| 118 |
|
| 119 |
\end{pud::man} |
| 120 |
|