Classify clouds in a model

Classify clouds in a model

Usage

classifyModel(mdata, mname, ttmx_dir, suffix = ".ttmx.dist.pac", k = 8)

Arguments

mdata

Data of a model, as given by summarizeHDBSCAN.

mname

Name of a model; the name of mdata.

ttmx_dir

Directory where the token-level distance matrices are stored.

suffix

Suffix of the file names for the token-level distance matrices; the function assumes that the name of the file is the name of the medoid plus the suffix.

k

Number of nearest neighbors for clusterDistance and clusterSeparation.

Value

A table with one row per cluster in the model, the columns created by clusterSeparation, clusterHDBSCAN and clusterDistance and the classification of each cluster based on the Nephological Shapes from Montes (2021) (see Chapter 5for a full description and examples).

References

Montes M (2021). Cloudspotting: Visual Analytics for Distributional Semantics. Ph.D. thesis, KU Leuven, Leuven.

Examples

if (FALSE) {
purrr::imap_dfr(models$medoidCoords, classifyModel, ttmx_dir = 'path/to/dir')
}