rustkernel-ml 0.2.0

Statistical ML kernels: clustering, anomaly detection, regression
Documentation

rustkernel-ml

Crates.io Documentation License

GPU-accelerated machine learning kernels for clustering, anomaly detection, regression, and explainability.

Kernels (17)

Clustering (3 kernels)

  • KMeans - Lloyd's algorithm with K-Means++ initialization
  • DBSCAN - Density-based clustering with GPU union-find
  • HierarchicalClustering - Agglomerative clustering

Anomaly Detection (4 kernels)

  • IsolationForest - Ensemble of isolation trees
  • LocalOutlierFactor - k-NN density estimation
  • StreamingIsolationForest - Online anomaly detection
  • AdaptiveThreshold - Self-adjusting thresholds with drift detection

Regression (2 kernels)

  • LinearRegression - OLS with regularization
  • LogisticRegression - Binary/multinomial classification

NLP & Embeddings (2 kernels)

  • EmbeddingGeneration - Hash-based text embeddings
  • SemanticSimilarity - Multi-metric similarity search

Healthcare (2 kernels)

  • DrugInteractionPrediction - Multi-drug interaction prediction
  • ClinicalPathwayConformance - Treatment guideline checking

Federated Learning (1 kernel)

  • SecureAggregation - Differential privacy aggregation

Explainability (2 kernels)

  • SHAPValues - Kernel SHAP for feature explanations
  • FeatureImportance - Permutation-based importance

Dimensionality Reduction (1 kernel)

  • PCA - Principal component analysis

Installation

Add to your Cargo.toml:

[dependencies]
rustkernel-ml = "0.1.0"

Usage

use rustkernel_ml::prelude::*;

// Create a KMeans kernel
let kmeans = KMeans::new();

// Cluster data points
let result = kmeans.cluster(&data, k, max_iterations);

License

Apache-2.0