MeanShiftOptions<T>

Options & Configuration · AiDotNet.Clustering.Options

Configuration options for Mean Shift clustering.

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

For Beginners

Mean Shift finds the “peaks” in your data.

Imagine your data as a landscape with hills:

  • Each point starts at its location
  • Points “roll uphill” toward the nearest peak
  • Points that end up at the same peak form a cluster

Key features:

  • Automatically determines number of clusters
  • Works well for finding dense regions
  • Good for image segmentation

Main parameter: bandwidth (how wide the “hills” are)

  • Small bandwidth: Many small clusters
  • Large bandwidth: Few large clusters

How It Works

Mean Shift is a non-parametric clustering algorithm that doesn’t require specifying the number of clusters. It finds clusters by iteratively shifting points toward the mode (densest area) of the local density.

Constructors

Constructor Summary
MeanShiftOptions()

Properties

Property Summary
NeighborAlgorithm Algorithm { get; set; } Gets or sets the neighbor finding algorithm.
bool BinSeeding { get; set; } Gets or sets whether to bin the seeds to speed up computation.
bool ClusterAll { get; set; } Gets or sets whether to cluster all points or just the seeds.
double BandwidthQuantile { get; set; } Gets or sets the quantile for automatic bandwidth estimation.
double? Bandwidth { get; set; } Gets or sets the bandwidth parameter.
double? ClusterMergeThreshold { get; set; } Gets or sets the minimum distance between cluster centers.
int LeafSize { get; set; } Gets or sets the leaf size for tree algorithms.
int MinBinFrequency { get; set; } Gets or sets the minimum number of points to consider a cluster.
int NumJobs { get; set; } Gets or sets the number of parallel jobs.