Clustering

Clustering inference assigns each new data point to one of the clusters discovered during training. Provide the same feature columns used during training — the model returns a cluster label for each row.

Available Models

• K-Means – Partition data into K spherical clusters
• DBSCAN – Density-based clustering with noise detection
• Hierarchical Clustering – Tree-based agglomerative clustering
• Mean Shift – Automatically discover cluster centroids from data density
• Gaussian Mixture Model – Soft probabilistic cluster assignments
• OPTICS – Density-based clustering robust to varying density
• Spectral Clustering – Graph-based clustering for non-convex shapes
• BIRCH – Scalable incremental clustering for large datasets
• Affinity Propagation – Message-passing algorithm, auto-selects cluster count
• Mini-Batch K-Means – Scalable K-Means using random mini-batches
• Bisecting K-Means – Hierarchical K-Means via recursive bisection