stream (version 1.5-0)

EvalCallback-class: Abstract Class for Evaluation Callbacks

Description

The abstract class for all evaluation callbacks. Cannot be instantiated. Must be inherited. Evaluation is the process of the clustering quality assessment. This assessment can include clustering results, as well as the clustering process, e.g., duration, spatial query performance, and similar. The stream package has some measurements (see evaluate for details) already implemented. All other measurements can be externally implemented without need to extend the stream package, by using callbacks.

Fields

all_measures

A list of all measures this object contributes to the evaluation. Union of all callback measures defines measures the end-user can use.

internal_measures

A list of internal measures. A subset of all_measures.

external_measures

A list of external measures. A subset of all_measures.

outlier_measures

A list of outlier measures. A subset of all_measures.

Methods

evaluate_callback(cb_obj, dsc, measure, points, actual, predict, outliers, predict_outliers, predict_outliers_corrid, centers, noise, ...)

A method that allows callback for external clustering results evaluation.

• cb_obj - The callback object (EvalCallback).

• dsc - The clusterer object (DSC).

• measure - The requested measures.

• points - A data frame contining all data items.

• actual - Actual assignments for the related data instance in points, given by the used data stream generator.

• predict - Assignments for the related data instance in points, given by the clusterer.

• outliers - Outlier marks for the related data instance in points, marked by the used data stream generator.

• predict_outliers - Outlier marks for the related data instance in points, marked by the clusterer.

• predict_outliers_corrid - Outlier identifiers assigned by the clusterer.

• centers - Cluster centers given by the clusterer.

• noise - Noise assignments (NA) for the related data instance in points, for all data instances that cannot be classified neither into clusters or outliers.

Examples

# NOT RUN {
CustomCallback <- function() {
  env <- environment()
  all_measures <- c("LowestWeightPercentage")
  internal_measures <- c()
  external_measures <- all_measures
  outlier_measures <- c()
  this <- list(description = "Custom evaluation callback",
               env = environment())
  class(this) <- c("CustomCallback", "EvalCallback")
  this
}
evaluate_callback.CustomCallback <- function(cb_obj, dsc, measure, points,
                                             actual, predict, outliers,
                                             predict_outliers,
                                             predict_outliers_corrid,
                                             centers, noise) {
    r <- list()
    if("LowestWeightPercentage" %in% measure)
        r$LowestWeightPercentage=min(get_weights(dsc))/sum(get_weights(dsc))
    r
}
stream <- DSD_Gaussians(k = 3, d = 2, p = c(0.2, 0.4, 0.4))
km <- DSC_Kmeans(3)
update(km, stream, n=500)
evaluate_with_callbacks(km, stream, type="macro", n=500,
                        measure = c("crand","LowestWeightPercentage"),
                        callbacks = list(cc=CustomCallback()))
# }