Merge pull request #65 from Arkoniak/ying_yang_algorithm

Implementation of Yinyang algorithm

Author: PyDataBlog

docs/src/index.md

@@ -56,11 +56,14 @@ git checkout experimental
 - [X] Implementation of [Hamerly implementation](https://www.researchgate.net/publication/220906984_Making_k-means_Even_Faster).
 - [X] Interface for inclusion in Alan Turing Institute's [MLJModels](https://github.com/alan-turing-institute/MLJModels.jl#who-is-this-repo-for).
 - [X] Full Implementation of Triangle inequality based on [Elkan - 2003 Using the Triangle Inequality to Accelerate K-Means"](https://www.aaai.org/Papers/ICML/2003/ICML03-022.pdf).
+- [X] Implementation of [Yinyang K-Means: A Drop-In Replacement of the Classic K-Means
+with Consistent Speedup](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/ding15.pdf)
 - [ ] Implementation of [Geometric methods to accelerate k-means algorithm](http://cs.baylor.edu/~hamerly/papers/sdm2016_rysavy_hamerly.pdf).
 - [ ] Support for other distance metrics supported by [Distances.jl](https://github.com/JuliaStats/Distances.jl#supported-distances).
 - [ ] Native support for tabular data inputs outside of MLJModels' interface.
 - [ ] Refactoring and finalizaiton of API desgin.
 - [ ] GPU support.
+- [ ] Distributed calculations support.
 - [ ] Implementation of other K-Means algorithm variants based on recent literature.
 - [ ] Optimization of code base.
 - [ ] Improved Documentation
@@ -103,6 +106,7 @@ r.converged             # whether the procedure converged
 - [Lloyd()](https://cs.nyu.edu/~roweis/csc2515-2006/readings/lloyd57.pdf)
 - [Hamerly()](https://www.researchgate.net/publication/220906984_Making_k-means_Even_Faster)
 - [Elkan()](https://www.aaai.org/Papers/ICML/2003/ICML03-022.pdf)
+- [Yinyang()](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/ding15.pdf)
 - [Geometric()](http://cs.baylor.edu/~hamerly/papers/sdm2016_rysavy_hamerly.pdf) - (Coming soon)
 - [MiniBatch()](https://www.eecs.tufts.edu/~dsculley/papers/fastkmeans.pdf) - (Coming soon)
 
@@ -174,8 +178,10 @@ ________________________________________________________________________________
 
 - 0.1.0 Initial release.
 - 0.1.1 Added interface for MLJ.
-- 0.1.2 Added Elkan algorithm.
+- 0.1.2 Added `Elkan` algorithm.
 - 0.1.3 Faster & optimized execution.
+- 0.1.4 Bug fixes
+- 0.1.5 Added `Yinyang` algorithm.
 
 ## Contributing
 

src/ParallelKMeans.jl

@@ -12,9 +12,10 @@ include("kmeans.jl")
 include("lloyd.jl")
 include("hamerly.jl")
 include("elkan.jl")
+include("yinyang.jl")
 include("mlj_interface.jl")
 
 export kmeans
-export Lloyd, Hamerly, Elkan
+export Lloyd, Hamerly, Elkan, Yinyang
 
 end # module

src/elkan.jl

@@ -236,21 +236,6 @@ function chunk_update_centroids(::Elkan, containers, centroids, X, r, idx)
     end
 end
 
-function collect_containers(alg::Elkan, containers, n_threads)
-    if n_threads == 1
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[1] ./ containers.centroids_cnt[1]'
-    else
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[1]
-        @inbounds containers.centroids_cnt[end] .= containers.centroids_cnt[1]
-        @inbounds for i in 2:n_threads
-            containers.centroids_new[end] .+= containers.centroids_new[i]
-            containers.centroids_cnt[end] .+= containers.centroids_cnt[i]
-        end
-
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[end] ./ containers.centroids_cnt[end]'
-    end
-end
-
 function calculate_centroids_movement(alg::Elkan, containers, centroids)
     p = containers.p
     centroids_new = containers.centroids_new[end]
@@ -260,7 +245,6 @@ function calculate_centroids_movement(alg::Elkan, containers, centroids)
     end
 end
 
-
 function chunk_update_bounds(alg, containers, centroids, r, idx)
     p = containers.p
     lb = containers.lb

src/hamerly.jl

@@ -73,23 +73,6 @@ function kmeans!(alg::Hamerly, containers, X, k;
     return KmeansResult(centroids, containers.labels, Float64[], Int[], Float64[], totalcost, niters, converged)
 end
 
-
-function collect_containers(alg::Hamerly, containers, n_threads)
-    if n_threads == 1
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[1] ./ containers.centroids_cnt[1]'
-    else
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[1]
-        @inbounds containers.centroids_cnt[end] .= containers.centroids_cnt[1]
-        @inbounds for i in 2:n_threads
-            containers.centroids_new[end] .+= containers.centroids_new[i]
-            containers.centroids_cnt[end] .+= containers.centroids_cnt[i]
-        end
-
-        @inbounds containers.centroids_new[end] .= containers.centroids_new[end] ./ containers.centroids_cnt[end]'
-    end
-end
-
-
 function create_containers(alg::Hamerly, k, nrow, ncol, n_threads)
     lng = n_threads + 1
     centroids_new = Vector{Array{Float64,2}}(undef, lng)
@@ -108,7 +91,7 @@ function create_containers(alg::Hamerly, k, nrow, ncol, n_threads)
 
     labels = zeros(Int, ncol)
 
-    # distance that centroid moved
+    # distance that centroid has moved
     p = Vector{Float64}(undef, k)
 
     # distance from the center to the closest other center
@@ -289,9 +272,9 @@ function chunk_update_bounds(alg::Hamerly, containers, r1, r2, pr1, pr2, r, idx)
         label = labels[i]
         ub[i] += 2*sqrt(abs(ub[i] * p[label])) + p[label]
         if r1 == label
-            lb[i] += pr2 - 2*sqrt(abs(pr2*lb[i]))
+            lb[i] = lb[i] <= pr2 ? 0.0 : lb[i] + pr2 - 2*sqrt(abs(pr2*lb[i]))
         else
-            lb[i] += pr1 - 2*sqrt(abs(pr1*lb[i]))
+            lb[i] = lb[i] <= pr1 ? 0.0 : lb[i] + pr1 - 2*sqrt(abs(pr1*lb[i]))
         end
     end
 end

src/kmeans.jl

@@ -159,3 +159,19 @@ function kmeans(alg, design_matrix, k;
                     k_init = k_init, max_iters = max_iters, tol = tol,
                     verbose = verbose, init = init)
 end
+
+
+function collect_containers(::AbstractKMeansAlg, containers, n_threads)
+    if n_threads == 1
+        @inbounds containers.centroids_new[end] .= containers.centroids_new[1] ./ containers.centroids_cnt[1]'
+    else
+        @inbounds containers.centroids_new[end] .= containers.centroids_new[1]
+        @inbounds containers.centroids_cnt[end] .= containers.centroids_cnt[1]
+        @inbounds for i in 2:n_threads
+            containers.centroids_new[end] .+= containers.centroids_new[i]
+            containers.centroids_cnt[end] .+= containers.centroids_cnt[i]
+        end
+
+        @inbounds containers.centroids_new[end] .= containers.centroids_new[end] ./ containers.centroids_cnt[end]'
+    end
+end