Compute embedding-based metrics — getEmbeddingMetrics • poem

Computes embedding-based metrics for the specified level.

Usage

getEmbeddingMetrics(
  x,
  labels,
  metrics = c("meanSW", "minSW", "pnSW", "dbcv"),
  distance = "euclidean",
  level = "class",
  ...
)

Arguments

x: A data.frame or matrix (with features as columns and items as rows) from which the metrics will be computed.
labels: A vector containing the labels of the predicted clusters. Must be a vector of characters, integers, numerics, or a factor, but not a list.
metrics: The metrics to compute. See details.
distance: The distance metric to use (default euclidean).
level: The level to calculate the metrics. Options include "element", "class" and "dataset".
...: Optional arguments. See details.

Value

A data.frame of metrics.

Details

The allowed values for metrics depend on the value of level:

If level = "element", the allowed metrics are: "SW".
If level = "class", the allowed metrics are: "meanSW", "minSW", "pnSW", "dbcv".
If level = "dataset", the allowed metrics are: "meanSW", "meanClassSW", "pnSW", "minClassSW", "cdbw", "cohesion", "compactness", "sep", "dbcv".

The function(s) that the optional arguments ... passed to depend on the value of level:

If level = "element", optional arguments are passed to stats::dist().
If level = "class", optional arguments are passed to dbcv().
If level = "dataset", optional arguments are passed to dbcv() or CDbw().

Examples

d1 <- mockData()
getEmbeddingMetrics(d1[,1:2], labels=d1$class, 
metrics=c("meanSW", "minSW", "pnSW", "dbcv"), level="class")
#>   class    meanSW      minSW pnSW       dbcv
#> 1     A 0.2398886 -0.3123599 0.12 -0.5434214
#> 2     B 0.1255021 -0.4781851 0.20 -0.6382858