deeptime.clustering

The clustering package contains clustering algorithms that can be used for discretization of time series.

Estimators

KMeans(n_clusters[, max_iter, metric, ...])	Clusters the data in a way that minimizes the cost function
MiniBatchKMeans(n_clusters[, batch_size, ...])	K-means clustering in a mini-batched fashion.
RegularSpace(dmin[, max_centers, metric, n_jobs])	Clusters data objects in such a way, that cluster centers are at least in distance of dmin to each other according to the given metric.
BoxDiscretization(dim, n_boxes[, v0, v1])	An n-dimensional box discretization of Euclidean space.

Models

ClusterModel(cluster_centers[, metric, ...])	A generic clustering model.
KMeansModel(cluster_centers, metric[, ...])	The K-means clustering model.
BoxDiscretizationModel(cluster_centers, v0, ...)	Model produced by BoxDiscretization.

Functions

kmeans_plusplus(data, n_clusters[, metric, ...])

Performs kmeans++ initialization.

Adding a new metric

metrics	Registry of available metrics.
MetricRegistry()	Registry of available metrics.