Skip to content
Snippets Groups Projects

Changes

Page history
  • New page
  • Templates
    • Update PrologMethods/Clustering/kmeans authored by Dean Samuel Schmitz's avatar Dean Samuel Schmitz
    Show whitespace changes
    Inline Side-by-side
    PrologMethods/Clustering/kmeans.md 0 → 100644
    View page @ c09ba4ff
    # K-Means Clustering
    An implementation of several strategies for efficient k-means clustering. Given a dataset and a value of k, this computes and returns a k-means clustering on that data.
    # Available Predicates
    * [naiveKMeans/12](/PrologMethods/Clustering/kmeans#naivekmeans12)
    * [dualTreeKMeans/12](/PrologMethods/Clustering/kmeans#dualtreekmeans12)
    * [elkanKMeans/12](/PrologMethods/Clustering/kmeans#elkankmeans12)
    * [hamerlyKMeans/12](/PrologMethods/Clustering/kmeans#hamerlykmeans12)
    * [pellegMooreKMeans/12](/PrologMethods/Clustering/kmeans#pellegmoorekmeans12)
    ---
    [links/resources](/PrologMethods/Clustering/kmeans#connected-linksresources)
    ## **_naiveKMeans/12_**
    Runs kmeans with naive as the algorithm for the Lloyd iteration.
    ```prolog
    %% part of the predicate definition
    naiveKMeans( +integer, +integer, +integer,
    +pointer(float_array), +integer, +integer,
    +integer,
    -pointer(float_array), -integer,
    -pointer(float_array), -integer, -integer).
    ```
    ### Parameters
    | Name | Type | Description | Default |
    |------|------|-------------|---------|
    | maxIterations | +integer | Maximum number of iterations before k-means terminates. | 1000 |
    | initialPartition | +string | Sets the initialPartitionPolicy : "SampleInitialzation", "RandomPartition" | SampleInitialzation |
    | emptyCluster | +string | Sets the emptyClusterPolicy: "MaxVarianceNewCluster", "KillEmptyCluster", "AllowEmptyCluster" | AllowEmptyCluster |
    | data | +matrix | Input dataset to perform clustering on. | - |
    | clusters | +integer | Number of clusters to find (0 autodetects from initial centroids). | 0 |
    | assignments | -vector | Vector to store cluster assignments in. | - |
    | centroids | -matrix | Matrix in which centroids are stored. | - |
    ---
    ## **_dualTreeKMeans/12_**
    Runs kmeans with dualtree as the algorithm for the Lloyd iteration.
    ```prolog
    %% part of the predicate definition
    dualTreeKMeans( +integer, +integer, +integer,
    +pointer(float_array), +integer, +integer,
    +integer,
    -pointer(float_array), -integer,
    -pointer(float_array), -integer, -integer).
    ```
    ### Parameters
    | Name | Type | Description | Default |
    |------|------|-------------|---------|
    | maxIterations | +integer | Maximum number of iterations before k-means terminates. | 1000 |
    | initialPartition | +string | Sets the initialPartitionPolicy : "SampleInitialzation", "RandomPartition" | SampleInitialzation |
    | emptyCluster | +string | Sets the emptyClusterPolicy: "MaxVarianceNewCluster", "KillEmptyCluster", "AllowEmptyCluster" | AllowEmptyCluster |
    | data | +matrix | Input dataset to perform clustering on. | - |
    | clusters | +integer | Number of clusters to find (0 autodetects from initial centroids). | 0 |
    | assignments | -vector | Vector to store cluster assignments in. | - |
    | centroids | -matrix | Matrix in which centroids are stored. | - |
    ---
    ## **_elkanKMeans/12_**
    Runs kmeans with elkan as the algorithm for the Lloyd iteration.
    ```prolog
    %% part of the predicate definition
    elkanKMeans( +integer, +integer, +integer,
    +pointer(float_array), +integer, +integer,
    +integer,
    -pointer(float_array), -integer,
    -pointer(float_array), -integer, -integer).
    ```
    ### Parameters
    | Name | Type | Description | Default |
    |------|------|-------------|---------|
    | maxIterations | +integer | Maximum number of iterations before k-means terminates. | 1000 |
    | initialPartition | +string | Sets the initialPartitionPolicy : "SampleInitialzation", "RandomPartition" | SampleInitialzation |
    | emptyCluster | +string | Sets the emptyClusterPolicy: "MaxVarianceNewCluster", "KillEmptyCluster", "AllowEmptyCluster" | AllowEmptyCluster |
    | data | +matrix | Input dataset to perform clustering on. | - |
    | clusters | +integer | Number of clusters to find (0 autodetects from initial centroids). | 0 |
    | assignments | -vector | Vector to store cluster assignments in. | - |
    | centroids | -matrix | Matrix in which centroids are stored. | - |
    ---
    ## **_hamerlyKMeans/12_**
    Runs kmeans with hamerly as the algorithm for the Lloyd iteration.
    ```prolog
    %% part of the predicate definition
    hamerlyKMeans( +integer, +integer, +integer,
    +pointer(float_array), +integer, +integer,
    +integer,
    -pointer(float_array), -integer,
    -pointer(float_array), -integer, -integer).
    ```
    ### Parameters
    | Name | Type | Description | Default |
    |------|------|-------------|---------|
    | maxIterations | +integer | Maximum number of iterations before k-means terminates. | 1000 |
    | initialPartition | +string | Sets the initialPartitionPolicy : "SampleInitialzation", "RandomPartition" | SampleInitialzation |
    | emptyCluster | +string | Sets the emptyClusterPolicy: "MaxVarianceNewCluster", "KillEmptyCluster", "AllowEmptyCluster" | AllowEmptyCluster |
    | data | +matrix | Input dataset to perform clustering on. | - |
    | clusters | +integer | Number of clusters to find (0 autodetects from initial centroids). | 0 |
    | assignments | -vector | Vector to store cluster assignments in. | - |
    | centroids | -matrix | Matrix in which centroids are stored. | - |
    ---
    ## **_pellegMooreKMeans/12\]_**
    Runs kmeans with pelleg morre as the algorithm for the Lloyd iteration.
    ```prolog
    %% part of the predicate definition
    pellegMooreKMeans( +integer, +integer, +integer,
    +pointer(float_array), +integer, +integer,
    +integer,
    -pointer(float_array), -integer,
    -pointer(float_array), -integer, -integer).
    ```
    ### Parameters
    | Name | Type | Description | Default |
    |------|------|-------------|---------|
    | maxIterations | +integer | Maximum number of iterations before k-means terminates. | 1000 |
    | initialPartition | +string | Sets the initialPartitionPolicy : "SampleInitialzation", "RandomPartition" | SampleInitialzation |
    | emptyCluster | +string | Sets the emptyClusterPolicy: "MaxVarianceNewCluster", "KillEmptyCluster", "AllowEmptyCluster" | AllowEmptyCluster |
    | data | +matrix | Input dataset to perform clustering on. | - |
    | clusters | +integer | Number of clusters to find (0 autodetects from initial centroids). | 0 |
    | assignments | -vector | Vector to store cluster assignments in. | - |
    | centroids | -matrix | Matrix in which centroids are stored. | - |
    ---
    # Connected Links/Resources
    If you want a more detailed explanation, then go to the python documentation. There is most of the time a good explanation on how the methods work and what the parameters do.
    * [MLpack::kmeans_C++\_documentation](https://www.mlpack.org/doc/stable/doxygen/classmlpack_1_1kmeans_1_1KMeans.html)
    * [MLpack::kmeans_Python_documentation](https://www.mlpack.org/doc/stable/python_documentation.html#kmeans)
    added some of the links from the python documentation
    * dbscan()
    * [Using the triangle inequality to accelerate k-means (pdf)](http://www.aaai.org/Papers/ICML/2003/ICML03-022.pdf)
    * [Making k-means even faster (pdf)](http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.586.2554&rep=rep1&type=pdf)
    * [Accelerating exact k-means algorithms with geometric reasoning (pdf)](http://reports-archive.adm.cs.cmu.edu/anon/anon/usr/ftp/usr0/ftp/2000/CMU-CS-00-105.pdf)
    * [A dual-tree algorithm for fast k-means clustering with large k (pdf)](http://www.ratml.org/pub/pdf/2017dual.pdf)
    \ No newline at end of file