|
MiniBatchKMeans Class |
Fast k-means clustering algorithm.
Inheritance HierarchySystemObject
Accord.MachineLearningParallelLearningBase
Accord.MachineLearningKMeans
Accord.MachineLearningMiniBatchKMeans
Accord.MachineLearningParallelLearningBase
Accord.MachineLearningKMeans
Accord.MachineLearningMiniBatchKMeans
Namespace: Accord.MachineLearning
Assembly: Accord.MachineLearning (in Accord.MachineLearning.dll) Version: 3.8.0
SyntaxThe MiniBatchKMeans type exposes the following members.
Constructors| Name | Description | |
|---|---|---|
 | MiniBatchKMeans(Int32, Int32) |
Initializes a new instance of KMeans algorithm
|
| MiniBatchKMeans(Int32, Int32, IDistanceDouble) |
Initializes a new instance of Mini-Batch K-Means algorithm
|
Properties| Name | Description | |
|---|---|---|
 | BatchSize |
Gets or sets the size of batches.
|
| Centroids |
Gets or sets the cluster centroids.
(Inherited from KMeans.) |
| Clusters |
Gets the clusters found by K-means.
(Inherited from KMeans.) |
| ComputeCovariances |
Gets or sets whether covariance matrices for the clusters should
be computed at the end of an iteration. Default is true.
(Inherited from KMeans.) |
| ComputeError |
Gets or sets whether the clustering distortion error (the
average distance between all data points and the cluster
centroids) should be computed at the end of the algorithm.
The result will be stored in Error. Default is true.
(Inherited from KMeans.) |
| Dimension |
Gets the dimensionality of the data space.
(Inherited from KMeans.) |
| Distance |
Gets or sets the distance function used
as a distance metric between data points.
(Inherited from KMeans.) |
| Error |
Gets the cluster distortion error after the
last call to this class' Compute methods.
(Inherited from KMeans.) |
| InitializationBatchSize |
Gets or sets the size of the batch used during initialization.
|
| Iterations |
Gets the number of iterations performed in the
last call to this class' Compute methods.
(Inherited from KMeans.) |
| K |
Gets the number of clusters.
(Inherited from KMeans.) |
| Labels |
Gets the labels assigned for each data point in the last
call to Learn(Double, Double).
|
| MaxIterations |
Gets or sets the maximum number of iterations to
be performed by the method. If set to zero, no
iteration limit will be imposed. Default is 0.
(Inherited from KMeans.) |
| NumberOfInitializations |
Gets or sets the number of different initializations of the centroids.
|
| ParallelOptions |
Gets or sets the parallelization options for this algorithm.
(Inherited from ParallelLearningBase.) |
| Token |
Gets or sets a cancellation token that can be used
to cancel the algorithm while it is running.
(Inherited from ParallelLearningBase.) |
| Tolerance |
Gets or sets the relative convergence threshold
for stopping the algorithm. Default is 1e-5.
(Inherited from KMeans.) |
| UseSeeding |
Gets or sets the strategy used to initialize the
centroids of the clustering algorithm. Default is
KMeansPlusPlus.
(Inherited from KMeans.) |
Methods| Name | Description | |
|---|---|---|
| Compute(Double) | Obsolete.
Divides the input data into K clusters.
(Inherited from KMeans.) |
| Compute(Double, Double) | Obsolete.
Divides the input data into K clusters.
(Inherited from KMeans.) |
| Compute(Double, Double) | Obsolete.
Divides the input data into K clusters.
(Inherited from KMeans.) |
 | ComputeInformation(Double) |
Computes the information about each cluster (covariance, proportions and error).
(Inherited from KMeans.) |
| ComputeInformation(Double, Int32) |
Computes the information about each cluster (covariance, proportions and error).
(Inherited from KMeans.) |
| converged |
Determines if the algorithm has converged by comparing the
centroids between two consecutive iterations.
(Inherited from KMeans.) |
| Equals | Determines whether the specified object is equal to the current object. (Inherited from Object.) |
| Finalize | Allows an object to try to free resources and perform other cleanup operations before it is reclaimed by garbage collection. (Inherited from Object.) |
| GetHashCode | Serves as the default hash function. (Inherited from Object.) |
| GetType | Gets the Type of the current instance. (Inherited from Object.) |
| Learn |
Learns a model that can map the given inputs to the desired outputs.
(Overrides KMeansLearn(Double, Double).) |
| MemberwiseClone | Creates a shallow copy of the current Object. (Inherited from Object.) |
| Randomize |
Randomizes the clusters inside a dataset.
(Inherited from KMeans.) |
| ToString | Returns a string that represents the current object. (Inherited from Object.) |
Extension Methods| Name | Description | |
|---|---|---|
 | HasMethod |
Checks whether an object implements a method with the given name.
(Defined by ExtensionMethods.) |
| IsEqual |
Compares two objects for equality, performing an elementwise
comparison if the elements are vectors or matrices.
(Defined by Matrix.) |
| To(Type) | Overloaded.
Converts an object into another type, irrespective of whether
the conversion can be done at compile time or not. This can be
used to convert generic types to numeric types during runtime.
(Defined by ExtensionMethods.) |
| ToT | Overloaded.
Converts an object into another type, irrespective of whether
the conversion can be done at compile time or not. This can be
used to convert generic types to numeric types during runtime.
(Defined by ExtensionMethods.) |
RemarksThe Mini-Batch K-Means clustering algorithm is a modification of the K-Means algorithm.
In each iteration, it uses only a portion of data to update the cluster centroids with the gradient step. The subsets of data are called mini-batches and are randomly sampled from the whole dataset in each iteration.
Mini-Batch K-Means is faster than k-means for large datasets since batching reduces computational time of the algorithm.
The algorithm is composed of the following steps:
- Place K points into the space represented by the objects that are being clustered. These points represent initial group centroids.
- Form a batch by choosing B objects from the whole input dataset. For each object in the batch, determine the group that has the closest centroid. Then, update the centroid with a gradient step.
- Repeat step 2 until the centroids converge or the maximal number of iterations has been performed.
References:
- D. Sculley. Web-Scale K-Means Clustering. Available on: https://www.eecs.tufts.edu/~dsculley/papers/fastkmeans.pdf
ExamplesAccord.Math.Random.Generator.Seed = 0; // Declare some observations double[][] observations = { new double[] { -5, -2, -1 }, new double[] { -5, -5, -6 }, new double[] { 2, 1, 1 }, new double[] { 1, 1, 2 }, new double[] { 1, 2, 2 }, new double[] { 3, 1, 2 }, new double[] { 11, 5, 4 }, new double[] { 15, 5, 6 }, new double[] { 10, 5, 6 }, }; // Create a new Mini-Batch K-Means algorithm MiniBatchKMeans mbkmeans = new MiniBatchKMeans(k: 3, batchSize: 2); // Compute and retrieve the data centroids var clusters = mbkmeans.Learn(observations); // Use the centroids to parition all the data int[] labels = clusters.Decide(observations);
See Also