Statistics::Gap 0.10

Statistics::Gap Perl module is an adaptation of the Gap Statistic.

SYNOPSIS

use Statistics::Gap;
$predictedk = &gap("prefix", "vec", INPUTMATRIX, "rbr", "h2", 30, 10, rep, 90, 4);

OR

use Statistics::Gap;
$predictedk = &gap("prefix", "vec", INPUTMATRIX, "rbr", "h2", 30, 10, rep, 90, 4, 7);

INPUTS

1. Prefix: The string that should be used to as a prefix while naming the intermediate files and the .dat files (plot files).
2. Space: Specifies the space in which the clustering should be performed. Valid parameter values: vec - vector space sim - similarity space
3. InputMatrix: Path to input matrix file. (More details about the input file-format below.)
4. ClusteringMethod: Specifies the clustering method to be used. (Learn more about this at: http://glaros.dtc.umn.edu/gkhome/cluto/cluto/overview)
Valid parameter values:
rb - Repeated Bisections
rbr - Repeated Bisections for by k-way refinement
direct - Direct k-way clustering
agglo - Agglomerative clustering
bagglo - Partitional biased Agglomerative clustering
NOTE: bagglo can be used only if space=vec
5. Crfun: Specifies the criterion function to be used for finding clustering solutions. (Learn more about this at: http://glaros.dtc.umn.edu/gkhome/cluto/cluto/overview)
Valid parameter values:
i1 - I1 Criterion function
i2 - I2 Criterion function
e1 - E1 Criterion function
h1 - H1 Criterion function
h2 - H2 Criterion function
6. K: This is an approximate upper bound for the number of clusters that may be present in the dataset.
7. B: The number of replicates/references to be generated.
8. TypeRef: Specifies whether to generate B replicates from a reference or to generate B references.
Valid parameter values:
rep - replicates
ref - references
9. Percentage: Specifies the percentage confidence to be reported in the log file. Since Statistics::Gap uses parametric bootstrap method for reference distribution generation, it is critical to understand the interval around the sample mean that could contain the population ("true") mean and with what certainty.
10. Precision: Specifies the precision to be used while generating the reference distribution.
11. Seed: The seed to be used with the random number generator. (This is an optional parameter. By default no seed is set.)