Crate scirs2_cluster 

SciRS2 Cluster - Clustering Algorithms

scirs2-cluster provides comprehensive clustering algorithms for unsupervised learning, offering k-means, hierarchical clustering, DBSCAN, spectral clustering, and advanced methods with parallel processing, SIMD acceleration, and evaluation metrics.

🎯 Key Features

• SciPy/scikit-learn Compatibility: Similar APIs to scipy.cluster and sklearn.cluster
• Partitional Clustering: K-means, K-means++, mini-batch K-means
• Hierarchical Clustering: Agglomerative with various linkage methods
• Density-based: DBSCAN, OPTICS, HDBSCAN for arbitrary-shaped clusters
• Graph-based: Spectral clustering, affinity propagation
• Evaluation Metrics: Silhouette, Davies-Bouldin, Calinski-Harabasz
• Performance: Parallel execution, SIMD distance computation

📦 Module Overview

SciRS2 Module	Python Equivalent	Description
vq	scipy.cluster.vq	K-means and vector quantization
hierarchy	scipy.cluster.hierarchy	Hierarchical/agglomerative clustering
dbscan	sklearn.cluster.DBSCAN	Density-based spatial clustering
spectral	sklearn.cluster.SpectralClustering	Graph-based spectral clustering
metrics	sklearn.metrics	Clustering evaluation metrics

🚀 Quick Start

[dependencies]
scirs2-cluster = "0.1.0-rc.2"

use scirs2_cluster::vq::kmeans;
use scirs2_core::ndarray::Array2;

// K-means clustering
let data = Array2::from_shape_vec((6, 2), vec![
    1.0, 2.0, 1.2, 1.8, 0.8, 1.9,
    3.7, 4.2, 3.9, 3.9, 4.2, 4.1,
]).unwrap();

let (centroids, labels) = kmeans(data.view(), 2, None, None, None, None).unwrap();

🔒 Version: 0.1.0-rc.2 (October 03, 2025)

Features

• Vector Quantization: K-means and K-means++ for partitioning data
• Hierarchical Clustering: Agglomerative clustering with various linkage methods
• Density-based Clustering: DBSCAN and OPTICS for finding clusters of arbitrary shape
• Mean Shift: Non-parametric clustering based on density estimation
• Spectral Clustering: Graph-based clustering using eigenvectors of the graph Laplacian
• Affinity Propagation: Message-passing based clustering that identifies exemplars
• Evaluation Metrics: Silhouette coefficient, Davies-Bouldin index, and other measures to evaluate clustering quality
• Data Preprocessing: Utilities for normalizing, standardizing, and whitening data before clustering

Examples

use scirs2_core::ndarray::{Array2, ArrayView2};
use scirs2_cluster::vq::kmeans;
use scirs2_cluster::preprocess::standardize;

// Example data with two clusters
let data = Array2::from_shape_vec((6, 2), vec![
    1.0, 2.0,
    1.2, 1.8,
    0.8, 1.9,
    3.7, 4.2,
    3.9, 3.9,
    4.2, 4.1,
]).unwrap();

// Standardize the data
let standardized = standardize(data.view(), true).unwrap();

// Run k-means with k=2
let (centroids, labels) = kmeans(standardized.view(), 2, None, None, None, None).unwrap();

// Print the results
println!("Centroids: {:?}", centroids);
println!("Cluster assignments: {:?}", labels);

Re-exports

pub use advanced::adaptive_online_clustering;

pub use advanced::deep_embedded_clustering;

pub use advanced::qaoa_clustering;

pub use advanced::quantum_kmeans;

pub use advanced::rl_clustering;

pub use advanced::transfer_learning_clustering;

pub use advanced::variational_deep_embedding;

pub use advanced::vqe_clustering;

pub use advanced::AdaptiveOnlineClustering;

pub use advanced::AdaptiveOnlineConfig;

pub use advanced::DeepClusteringConfig;

pub use advanced::DeepEmbeddedClustering;

pub use advanced::FeatureAlignment;

pub use advanced::QAOAClustering;

pub use advanced::QAOAConfig;

pub use advanced::QAOACostFunction;

pub use advanced::QuantumConfig;

pub use advanced::QuantumKMeans;

pub use advanced::RLClustering;

pub use advanced::RLClusteringConfig;

pub use advanced::RewardFunction;

pub use advanced::TransferLearningClustering;

pub use advanced::TransferLearningConfig;

pub use advanced::VQEAnsatz;

pub use advanced::VQEClustering;

pub use advanced::VQEConfig;

pub use advanced::VariationalDeepEmbedding;

pub use quantum_clustering::quantum_annealing_clustering;

pub use quantum_clustering::CoolingSchedule;

pub use quantum_clustering::QuantumAnnealingClustering;

pub use quantum_clustering::QuantumAnnealingConfig;

pub use advanced_clustering::AdvancedClusterer;

pub use advanced_clustering::AdvancedClusteringResult;

pub use advanced_clustering::AdvancedConfig;

pub use advanced_clustering::AdvancedPerformanceMetrics;

pub use advanced_visualization::create_advanced_visualization_report;

pub use advanced_visualization::visualize_advanced_results;

pub use advanced_visualization::AISelectionPlot;

pub use advanced_visualization::AdvancedVisualizationConfig;

pub use advanced_visualization::AdvancedVisualizationOutput;

pub use advanced_visualization::AdvancedVisualizer;

pub use advanced_visualization::ClusterPlot;

pub use advanced_visualization::NeuromorphicAdaptationPlot;

pub use advanced_visualization::PerformanceDashboard;

pub use advanced_visualization::QuantumCoherencePlot;

pub use advanced_visualization::QuantumColorScheme;

pub use advanced_visualization::VisualizationExportFormat;

pub use enhanced_clustering_features::DeepAdvancedClusterer;

pub use enhanced_clustering_features::DeepAdvancedResult;

pub use enhanced_clustering_features::DeepEnsembleCoordinator;

pub use enhanced_clustering_features::EnsembleConsensus;

pub use enhanced_clustering_features::GraphNeuralNetworkProcessor;

pub use enhanced_clustering_features::GraphStructureInsights;

pub use enhanced_clustering_features::NeuralArchitectureSearchEngine;

pub use enhanced_clustering_features::OptimalArchitecture;

pub use enhanced_clustering_features::ReinforcementLearningAgent;

pub use enhanced_clustering_features::SpectralProperties;

pub use enhanced_clustering_features::TransformerClusterEmbedder;

pub use gpu_distributed_clustering::CommunicationOverhead;

pub use gpu_distributed_clustering::CoordinationStrategy;

pub use gpu_distributed_clustering::DistributedAdvancedClusterer;

pub use gpu_distributed_clustering::DistributedAdvancedResult;

pub use gpu_distributed_clustering::DistributedProcessingMetrics;

pub use gpu_distributed_clustering::GpuAccelerationConfig;

pub use gpu_distributed_clustering::GpuAccelerationMetrics;

pub use gpu_distributed_clustering::GpuAdvancedClusterer;

pub use gpu_distributed_clustering::GpuAdvancedResult;

pub use gpu_distributed_clustering::GpuDeviceSelection;

pub use gpu_distributed_clustering::GpuMemoryStrategy;

pub use gpu_distributed_clustering::GpuOptimizationLevel;

pub use gpu_distributed_clustering::HybridGpuDistributedClusterer;

pub use gpu_distributed_clustering::HybridGpuDistributedResult;

pub use gpu_distributed_clustering::LoadBalancingStats;

pub use gpu_distributed_clustering::WorkerNodeConfig;

pub use gpu_distributed_clustering::WorkerPerformanceStats;

pub use advanced_benchmarking::create_comprehensive_report;

pub use advanced_benchmarking::AdvancedBenchmark;

pub use advanced_benchmarking::AlgorithmBenchmark;

pub use advanced_benchmarking::AlgorithmComparison;

pub use advanced_benchmarking::BenchmarkConfig;

pub use advanced_benchmarking::BenchmarkResults;

pub use advanced_benchmarking::ComplexityClass;

pub use advanced_benchmarking::GpuVsCpuComparison;

pub use advanced_benchmarking::MemoryProfile;

pub use advanced_benchmarking::OptimizationCategory;

pub use advanced_benchmarking::OptimizationPriority;

pub use advanced_benchmarking::OptimizationSuggestion;

pub use advanced_benchmarking::PerformanceStatistics;

pub use advanced_benchmarking::QualityMetrics;

pub use advanced_benchmarking::RegressionAlert;

pub use advanced_benchmarking::RegressionSeverity;

pub use advanced_benchmarking::ScalabilityAnalysis;

pub use advanced_benchmarking::SystemInfo;

pub use affinity::affinity_propagation;

pub use affinity::AffinityPropagationOptions;

pub use birch::birch;

pub use birch::Birch;

pub use birch::BirchOptions;

pub use birch::BirchStatistics;

pub use density::hdbscan::dbscan_clustering;

pub use density::hdbscan::hdbscan;

pub use density::hdbscan::ClusterSelectionMethod;

pub use density::hdbscan::HDBSCANOptions;

pub use density::hdbscan::HDBSCANResult;

pub use density::hdbscan::StoreCenter;

pub use density::optics::extract_dbscan_clustering;

pub use density::optics::extract_xi_clusters;

pub use density::optics::OPTICSResult;

pub use ensemble::convenience::bootstrap_ensemble;

pub use ensemble::convenience::ensemble_clustering;

pub use ensemble::convenience::multi_algorithm_ensemble;

pub use ensemble::ClusteringAlgorithm;

pub use ensemble::ClusteringResult;

pub use ensemble::ConsensusMethod;

pub use ensemble::ConsensusStatistics;

pub use ensemble::DiversityMetrics;

pub use ensemble::DiversityStrategy;

pub use ensemble::EnsembleClusterer;

pub use ensemble::EnsembleConfig;

pub use ensemble::EnsembleResult;

pub use ensemble::NoiseType;

pub use ensemble::ParameterRange;

pub use ensemble::SamplingStrategy;

pub use gmm::gaussian_mixture;

pub use gmm::CovarianceType;

pub use gmm::GMMInit;

pub use gmm::GMMOptions;

pub use gmm::GaussianMixture;

pub use graph::girvan_newman;

pub use graph::graph_clustering;

pub use graph::label_propagation;

pub use graph::louvain;

pub use graph::Graph;

pub use graph::GraphClusteringAlgorithm;

pub use graph::GraphClusteringConfig;

pub use input_validation::check_duplicate_points;

pub use input_validation::suggest_clustering_algorithm;

pub use input_validation::validate_clustering_data;

pub use input_validation::validate_convergence_parameters;

pub use input_validation::validate_distance_parameter;

pub use input_validation::validate_integer_parameter;

pub use input_validation::validate_n_clusters;

pub use input_validation::validate_sample_weights;

pub use input_validation::ValidationConfig;

pub use leader::euclidean_distance;

pub use leader::leader_clustering;

pub use leader::manhattan_distance;

pub use leader::LeaderClustering;

pub use leader::LeaderNode;

pub use leader::LeaderTree;

pub use meanshift::estimate_bandwidth;

pub use meanshift::get_bin_seeds;

pub use meanshift::mean_shift;

pub use meanshift::MeanShift;

pub use meanshift::MeanShiftOptions;

pub use metrics::adjusted_rand_index;

pub use metrics::calinski_harabasz_score;

pub use metrics::davies_bouldin_score;

pub use metrics::homogeneity_completeness_v_measure;

pub use metrics::normalized_mutual_info;

pub use metrics::silhouette_samples;

pub use metrics::silhouette_score;

pub use metrics::bootstrap_confidence_interval;

pub use metrics::information_theoretic::normalized_variation_of_information;

pub use metrics::jensen_shannon_divergence;

pub use metrics::advanced::bic_score;

pub use metrics::advanced::dunn_index;

pub use neighbor_search::create_neighbor_searcher;

pub use neighbor_search::BallTree;

pub use neighbor_search::BruteForceSearch;

pub use neighbor_search::KDTree;

pub use neighbor_search::NeighborResult;

pub use neighbor_search::NeighborSearchAlgorithm;

pub use neighbor_search::NeighborSearchConfig;

pub use neighbor_search::NeighborSearcher;

pub use preprocess::min_max_scale;

pub use preprocess::normalize;

pub use preprocess::standardize;

pub use preprocess::whiten;

pub use preprocess::NormType;

pub use serialization::affinity_propagation_to_model;

pub use serialization::birch_to_model;

pub use serialization::compatibility;

pub use serialization::dbscan_to_model;

pub use serialization::gmm_to_model;

pub use serialization::hierarchy_to_model;

pub use serialization::kmeans_to_model;

pub use serialization::leader_to_model;

pub use serialization::leadertree_to_model;

pub use serialization::meanshift_to_model;

pub use serialization::save_affinity_propagation;

pub use serialization::save_birch;

pub use serialization::save_gmm;

pub use serialization::save_hierarchy;

pub use serialization::save_kmeans;

pub use serialization::save_leader;

pub use serialization::save_leadertree;

pub use serialization::save_spectral_clustering;

pub use serialization::spectral_clustering_to_model;

pub use serialization::AdvancedExport;

pub use serialization::AffinityPropagationModel;

pub use serialization::AlgorithmState;

pub use serialization::AutoSaveConfig;

pub use serialization::BirchModel;

pub use serialization::ClusteringWorkflow;

pub use serialization::ClusteringWorkflowManager;

pub use serialization::DBSCANModel;

pub use serialization::DataCharacteristics;

pub use serialization::EnhancedModel;

pub use serialization::EnhancedModelMetadata;

pub use serialization::ExportFormat;

pub use serialization::GMMModel;

pub use serialization::HierarchicalModel;

pub use serialization::KMeansModel;

pub use serialization::LeaderModel;

pub use serialization::LeaderTreeModel;

pub use serialization::MeanShiftModel;

pub use serialization::ModelMetadata;

pub use serialization::PlatformInfo;

pub use serialization::SerializableModel;

pub use serialization::SpectralClusteringModel;

pub use serialization::TrainingMetrics;

pub use serialization::TrainingStep;

pub use serialization::WorkflowConfig;

pub use serialization::compatibility::create_sklearn_param_grid;

pub use serialization::compatibility::from_joblib_format;

pub use serialization::compatibility::from_numpy_format;

pub use serialization::compatibility::from_sklearn_format;

pub use serialization::compatibility::generate_sklearn_model_summary;

pub use serialization::compatibility::to_arrow_schema;

pub use serialization::compatibility::to_huggingface_card;

pub use serialization::compatibility::to_joblib_format;

pub use serialization::compatibility::to_mlflow_format;

pub use serialization::compatibility::to_numpy_format;

pub use serialization::compatibility::to_onnx_metadata;

pub use serialization::compatibility::to_pandas_clustering_report;

pub use serialization::compatibility::to_pandas_format;

pub use serialization::compatibility::to_pickle_like_format;

pub use serialization::compatibility::to_pytorch_checkpoint;

pub use serialization::compatibility::to_r_format;

pub use serialization::compatibility::to_scipy_dendrogram_format;

pub use serialization::compatibility::to_scipy_linkage_format;

pub use serialization::compatibility::to_sklearn_clustering_result;

pub use serialization::compatibility::to_sklearn_format;

pub use sparse::sparse_epsilon_graph;

pub use sparse::sparse_knn_graph;

pub use sparse::SparseDistanceMatrix;

pub use sparse::SparseHierarchicalClustering;

pub use spectral::spectral_bipartition;

pub use spectral::spectral_clustering;

pub use spectral::AffinityMode;

pub use spectral::SpectralClusteringOptions;

pub use stability::BootstrapValidator;

pub use stability::ConsensusClusterer;

pub use stability::OptimalKSelector;

pub use stability::StabilityConfig;

pub use stability::StabilityResult;

pub use streaming::ChunkedDistanceMatrix;

pub use streaming::ProgressiveHierarchical;

pub use streaming::StreamingConfig;

pub use streaming::StreamingKMeans;

pub use text_clustering::semantic_hierarchical;

pub use text_clustering::semantic_kmeans;

pub use text_clustering::topic_clustering;

pub use text_clustering::SemanticClusteringConfig;

pub use text_clustering::SemanticHierarchical;

pub use text_clustering::SemanticKMeans;

pub use text_clustering::SemanticSimilarity;

pub use text_clustering::TextPreprocessing;

pub use text_clustering::TextRepresentation;

pub use text_clustering::TopicBasedClustering;

pub use time_series::dtw_barycenter_averaging;

pub use time_series::dtw_distance;

pub use time_series::dtw_distance_custom;

pub use time_series::dtw_hierarchical_clustering;

pub use time_series::dtw_k_means;

pub use time_series::dtw_k_medoids;

pub use time_series::soft_dtw_distance;

pub use time_series::time_series_clustering;

pub use time_series::TimeSeriesAlgorithm;

pub use time_series::TimeSeriesClusteringConfig;

pub use tuning::AcquisitionFunction;

pub use tuning::AutoTuner;

pub use tuning::BayesianState;

pub use tuning::CVStrategy;

pub use tuning::ConvergenceInfo;

pub use tuning::CrossValidationConfig;

pub use tuning::EarlyStoppingConfig;

pub use tuning::EnsembleResults;

pub use tuning::EvaluationMetric;

pub use tuning::EvaluationResult;

pub use tuning::ExplorationStats;

pub use tuning::HyperParameter;

pub use tuning::KernelType;

pub use tuning::LoadBalancingStrategy;

pub use tuning::ParallelConfig;

pub use tuning::ResourceConstraints;

pub use tuning::SearchSpace;

pub use tuning::SearchStrategy;

pub use tuning::StandardSearchSpaces;

pub use tuning::StoppingReason;

pub use tuning::SurrogateModel;

pub use tuning::TuningConfig;

pub use tuning::TuningResult;

pub use visualization::create_scatter_plot_2d;

pub use visualization::create_scatter_plot_3d;

pub use visualization::AnimationConfig;

pub use visualization::BoundaryType;

pub use visualization::ClusterBoundary;

pub use visualization::ColorScheme;

pub use visualization::DimensionalityReduction;

pub use visualization::EasingFunction;

pub use visualization::LegendEntry;

pub use visualization::ScatterPlot2D;

pub use visualization::ScatterPlot3D;

pub use visualization::VisualizationConfig;

pub use visualization::animation::AnimationFrame;

pub use visualization::animation::IterativeAnimationConfig;

pub use visualization::animation::IterativeAnimationRecorder;

pub use visualization::animation::StreamingVisualizer;

pub use visualization::interactive::ClusterStats;

pub use visualization::interactive::InteractiveConfig;

pub use visualization::interactive::InteractiveState;

pub use visualization::interactive::InteractiveVisualizer;

pub use visualization::export::export_scatter_2d_to_html;

pub use visualization::export::export_scatter_2d_to_json;

pub use visualization::export::export_scatter_3d_to_html;

pub use visualization::export::export_scatter_3d_to_json;

pub use visualization::export::save_visualization_to_file;

pub use distributed::DataPartition;

pub use distributed::DistributedKMeans;

pub use distributed::DistributedKMeansConfig;

pub use distributed::PartitioningStrategy;

pub use distributed::WorkerStatus;

pub use density::*;

pub use hierarchy::*;

pub use vq::*;

Modules

advanced

Cutting-edge clustering algorithms including quantum-inspired methods and advanced online learning.

advanced_benchmarking

Advanced benchmarking and performance profiling system.

advanced_clustering

Advanced Clustering - AI-Driven Quantum-Neuromorphic Clustering

advanced_visualization

Enhanced visualization specifically for advanced clustering results.

affinity

Affinity Propagation clustering implementation

birch

BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies) clustering algorithm

density

Density-based clustering algorithms

distributed

Distributed clustering algorithms for large-scale datasets.

enhanced_clustering_features

Enhanced Advanced Features - Advanced AI-Driven Clustering Extensions

ensemble

Ensemble clustering methods for improved robustness.

error

Error types for the clustering module

gmm

Gaussian Mixture Models (GMM) for clustering

gpu_distributed_clustering

Advanced GPU and Distributed Computing Extensions

graph

Graph clustering and community detection algorithms.

hierarchy

Hierarchical clustering functions

input_validation

Enhanced input validation utilities

leader

Leader algorithm implementation for clustering

meanshift

Mean Shift clustering implementation.

metrics

Clustering evaluation metrics

neighbor_search

Efficient neighbor search algorithms for clustering

preprocess

Data preprocessing utilities for clustering algorithms

quantum_clustering

Quantum-Inspired Clustering Algorithms

serialization

Model serialization and deserialization

sparse

Sparse distance matrix support for large datasets

spectral

Spectral clustering implementation

stability

Cluster stability assessment tools

streaming

Streaming and memory-efficient clustering algorithms

text_clustering

Text clustering algorithms with semantic similarity support.

time_series

Time series clustering algorithms with specialized distance metrics.

tuning

Automatic hyperparameter tuning for clustering algorithms.

utils

Utility modules for clustering algorithms

visualization

Enhanced visualization capabilities for clustering results.

vq

Vector quantization functions