DenclueOptions<T>

Options & Configuration · AiDotNet.Clustering.Options

Configuration options for DENCLUE clustering.

Inheritance: ModelOptions → ClusteringOptions<T> → DenclueOptions<T>

For Beginners

DENCLUE sees data as a “landscape” where dense areas form “mountains” (attractors).

Imagine pouring water on a terrain:

  • Water flows downhill to valleys
  • In DENCLUE, points “flow uphill” to peaks
  • Points reaching the same peak form a cluster

Key parameters:

  • Bandwidth (h): Controls how wide the “mountains” are
  • Small h: Many narrow peaks (more clusters)
  • Large h: Few wide peaks (fewer clusters)
  • MinDensity: Minimum density for a peak to be a cluster center

How It Works

DENCLUE (DENsity-based CLUstEring) uses kernel density estimation to find clusters. Points are attracted to density maxima using gradient ascent, and clusters are formed by grouping points that converge to the same attractor.

Constructors

Constructor Summary
DenclueOptions() Initializes DenclueOptions with appropriate defaults.

Properties

Property Summary
double AttractorMergeThreshold { get; set; } Gets or sets the distance threshold for merging attractors.
double Bandwidth { get; set; } Gets or sets the bandwidth parameter for the Gaussian kernel.
double ConvergenceThreshold { get; set; } Gets or sets the convergence threshold for hill climbing.
double MinDensity { get; set; } Gets or sets the minimum density threshold for cluster attractors.