clmimac(1)

Interpret Matrices (c.q. MCL iterands output by mcl) As Clusterings.

Section 1 mcl bookworm source

Description

clm imac

NAME

clm_imac - Interpret Matrices (c.q. MCL iterands output by mcl) As Clusterings.

clmimac is not in actual fact a program. This manual page documents the behaviour and options of the clm program when invoked in mode imac. The options -h, --apropos, --version, -set, --nop are accessible in all clm modes. They are described in the clm manual page.

SYNOPSIS

clm imac -imx <fname> [options]

clm imac -imx fname (input file) [-o fname (name/stem for output)] [-dag fname (output DAG)] [-overlap str (overlap mode)] [-sort str (size|revsize|lex|none)] [-strict num (in 0..1)] [-h (print synopsis, exit)] [--apropos (print synopsis, exit)] [--version (print version, exit)]

DESCRIPTION

Use clm imac to interpret matrices (as clusterings) output by mcl using mcl’s -dump ite option.

Use clm imac only if you have a special reason; the normal usage of mcl is to do multiple runs for varying -I parameters and use the clusterings output by mcl itself. One reason is if you are interested in clusterings with overlap; early MCL iterands generally induce clusterings possessing overlap. Another reason is to investigate how the cluster structure associated with the MCL process evolves over time.

OPTIONS

-imx fname (input file)
The input file is presumably an MCL iterand resulting from the mcl option -dump ite.

-strict num (in 0..1)
Higher values (up until 1) will thin out the DAG constructed by clm imac. The default value is 0.00001, yielding the full DAG.

-o fname (file name/stem)
Write to file named fname.

-dag fname (output DAG)
Write the DAG (directed acyclic graph) constructed from the input to file. This DAG is constructed according to the structure associated with diagonally positive semi-definite matrices as described in the PhD thesis Graph clustering by flow simulation. Consult mclfamily(7) for references.

-sort str (size|revsize|lex|none)
Sort the clusters either by increasing size, decreasing size, lexicographically by the indices they contain, or use the clustering exactly as obtained from the interpretation routine.

-overlap mode (overlap mode)
With mode set to cut, remove any overlap by allocating the nodes in overlap to the first cluster in which they were found. Mode keep leaves overlap unchanged, and mode split results in overlapping parts excised and introduced as clusters in their own right.

AUTHOR

Stijn van Dongen.