version 1.1.0

DESCRIPTION

@@ -2,8 +2,8 @@ Package: ClusterR
 Type: Package
 Title: Gaussian Mixture Models, K-Means, Mini-Batch-Kmeans and
         K-Medoids Clustering
-Version: 1.0.9
-Date: 2017-11-30
+Version: 1.1.0
+Date: 2018-01-16
 Author: Lampros Mouselimis <mouselimislampros@gmail.com>
 Maintainer: Lampros Mouselimis <mouselimislampros@gmail.com>
 BugReports: https://github.com/mlampros/ClusterR/issues
@@ -22,6 +22,6 @@ Suggests: testthat, covr, knitr, rmarkdown
 VignetteBuilder: knitr
 RoxygenNote: 6.0.1
 NeedsCompilation: yes
-Packaged: 2017-11-30 12:02:51 UTC; lampros
+Packaged: 2018-01-16 15:22:47 UTC; lampros
 Repository: CRAN
-Date/Publication: 2017-11-30 13:00:05 UTC
+Date/Publication: 2018-01-17 18:01:58 UTC

MD5

@@ -1,23 +1,23 @@
-4ba5c0512cea5b202d6aa4865b8d3d48 *DESCRIPTION
+ace8b8631582c6ecac17bdd2655da786 *DESCRIPTION
 c9c69a326502c31a65128654c3c08cb5 *LICENSE
 1cf3e5d5262851d34692258a58312008 *NAMESPACE
-76a37150f8b9bd10d49f44cebc4a1570 *NEWS.md
+a526b34b4bf940e7fec070072b6e2d0c *NEWS.md
 19f3ce55803fd8c849bdfa8fd3cc831b *R/RcppExports.R
-afcc201781a94cc54334bf81ad399e07 *R/clustering_functions.R
+7baeb6643966f872dd5ee7f39fc03149 *R/clustering_functions.R
 ede10384907019d8669da63646557e31 *README.md
-c752579b23f60b5b51aa1e39e309fffc *build/vignette.rds
+e037a0892dccbc8c11badfea2e787664 *build/vignette.rds
 8b6687f3bb9c58cd74ae67670872215d *data/dietary_survey_IBS.rda
 aa58963ebd13c4c91edf809ca4efc5d4 *data/mushroom.rda
 a9e3dfd8650ed7d2d0a91d3880a67f7b *data/soybean.rda
 128fe5f74b8c6787e9c378b61bd8d423 *inst/doc/the_clusterR_package.R
 86f33097b23d7a5acc0c83647a19d865 *inst/doc/the_clusterR_package.Rmd
-bee08e59baf7f8da5a698c0607cae625 *inst/doc/the_clusterR_package.html
+2891d5475073f815afc44c8b859ed5a1 *inst/doc/the_clusterR_package.html
 94cd0d57e5b44e47a3746a7979a4088f *man/Clara_Medoids.Rd
 c4f35a38b1a4271caf0a29ed616c45ba *man/Cluster_Medoids.Rd
 853d94cb9bf0db952fee7ce47788baa5 *man/GMM.Rd
 add088ce3a97143c6a7920c2a5d04df4 *man/KMeans_arma.Rd
 bb8de9fbba458b06aff94061d2d3765e *man/KMeans_rcpp.Rd
-66152d79ea97f9fc524403e8a8f4acc6 *man/MiniBatchKmeans.Rd
+2219970918b3d69ad736f506c7b356b1 *man/MiniBatchKmeans.Rd
 e0427f0525c0407bbdcb48ca5dc8f7ca *man/Optimal_Clusters_GMM.Rd
 a0641377964d9bb3e2045614b50ffbd5 *man/Optimal_Clusters_KMeans.Rd
 c84b2645d14b6c057736ea9015fcc582 *man/Optimal_Clusters_Medoids.Rd
@@ -38,11 +38,11 @@ cd805bfc471c3894870346627bb3b047 *man/predict_Medoids.Rd
 6c290a2ae91a58e535ef1adac77e8e57 *man/tryCatch_GMM.Rd
 9e21127499fb7ce22636287667ea3b34 *man/tryCatch_KMEANS_arma.Rd
 62231fa9f6be487bdbf9f95625d39c0b *man/tryCatch_optimal_clust_GMM.Rd
-16d006bf40b078243e18207776416cf7 *src/Makevars
-18579fb24623b1813f3972b33fe55dac *src/Makevars.win
+c0b0887d0e21dba427791387a69357a4 *src/Makevars
+3924f33984346427452b9a43de77a4c8 *src/Makevars.win
 ba1dafe01a73b757ea17e6c04dee7b39 *src/RcppExports.cpp
 8be2c73e7a0e14a4500c68d9444cc5cd *src/init.c
-0ba1c6644b1357bcce0fcb3aeb604886 *src/kmeans_miniBatchKmeans_GMM_Medoids.cpp
+02132d8c61a2e864cbe93d557dd70829 *src/kmeans_miniBatchKmeans_GMM_Medoids.cpp
 580cce096456bdd88fc875489a33be99 *src/utils_rcpp.cpp
 76f46da484cec9e5763f21f16c309c50 *src/utils_rcpp.h
 3243b3e7b85ca7953f191679629429ec *tests/testthat.R

NEWS.md

@@ -1,4 +1,10 @@
 
+## ClusterR 1.1.0
+
+* I added the *DARMA_64BIT_WORD* flag in the Makevars file to allow the package processing big datasets
+* I modified the *kmeans_miniBatchKmeans_GMM_Medoids.cpp* file and especially all *Rcpp::List::create()* objects to addrress the clang-ASAN errors.
+
+
 ## ClusterR 1.0.9
 
 * I modified the *Optimal_Clusters_KMeans* function to return a vector with the *distortion_fK* values if criterion is *distortion_fK* (instead of the *WCSSE* values).

R/clustering_functions.R

@@ -811,7 +811,7 @@ Optimal_Clusters_KMeans = function(data, max_clusters, criterion = "variance_exp
 #' \strong{random}         : random selection of data rows as initial centroids
 #'
 #' @references
-#' http://www.eecs.tufts.edu/~dsculley/papers/fastkmeans.pdf
+#' http://www.eecs.tufts.edu/~dsculley/papers/fastkmeans.pdf, https://github.com/siddharth-agrawal/Mini-Batch-K-Means
 #' @export
 #' @examples
 #'

inst/doc/the_clusterR_package.html

@@ -12,7 +12,7 @@
 
 <meta name="author" content="Lampros Mouselimis" />
 
-<meta name="date" content="2017-11-30" />
+<meta name="date" content="2018-01-16" />
 
 <title>Functionality of the ClusterR package</title>
 
@@ -70,7 +70,7 @@
 
 <h1 class="title toc-ignore">Functionality of the ClusterR package</h1>
 <h4 class="author"><em>Lampros Mouselimis</em></h4>
-<h4 class="date"><em>2017-11-30</em></h4>
+<h4 class="date"><em>2018-01-16</em></h4>
 
 
 
@@ -258,7 +258,7 @@ <h5>Mini-batch-kmeans</h5>
 t =<span class="st"> </span>end <span class="op">-</span><span class="st"> </span>start
 
 <span class="kw">cat</span>(<span class="st">'time to complete :'</span>, t, <span class="kw">attributes</span>(t)<span class="op">$</span>units, <span class="st">'</span><span class="ch">\n</span><span class="st">'</span>)</code></pre></div>
-<pre><code>## time to complete : 2.275446 secs</code></pre>
+<pre><code>## time to complete : 2.275485 secs</code></pre>
 <div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r">getcent_init =<span class="st"> </span>km_init<span class="op">$</span>centroids
 
 getclust_init =<span class="st"> </span>km_init<span class="op">$</span>clusters
@@ -286,7 +286,7 @@ <h5>Mini-batch-kmeans</h5>
 t =<span class="st"> </span>end <span class="op">-</span><span class="st"> </span>start
 
 <span class="kw">cat</span>(<span class="st">'time to complete :'</span>, t, <span class="kw">attributes</span>(t)<span class="op">$</span>units, <span class="st">'</span><span class="ch">\n</span><span class="st">'</span>)</code></pre></div>
-<pre><code>## time to complete : 0.8078771 secs</code></pre>
+<pre><code>## time to complete : 0.8378835 secs</code></pre>
 <div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r">getcent_mb =<span class="st"> </span>km_mb<span class="op">$</span>centroids
 
 new_im_mb =<span class="st"> </span>getcent_mb[pr_mb, ]   <span class="co"># each observation is associated with the nearby centroid</span>
@@ -490,7 +490,7 @@ <h5>K-Medoids</h5>
 t =<span class="st"> </span>end <span class="op">-</span><span class="st"> </span>start
 
 <span class="kw">cat</span>(<span class="st">'time to complete :'</span>, t, <span class="kw">attributes</span>(t)<span class="op">$</span>units, <span class="st">'</span><span class="ch">\n</span><span class="st">'</span>)</code></pre></div>
-<pre><code>## time to complete : 2.711706 secs</code></pre>
+<pre><code>## time to complete : 3.194762 secs</code></pre>
 <p><br></p>
 <table>
 <caption>hamming-Clara-Medoids</caption>
@@ -522,7 +522,7 @@ <h5>K-Medoids</h5>
 t =<span class="st"> </span>end <span class="op">-</span><span class="st"> </span>start
 
 <span class="kw">cat</span>(<span class="st">'time to complete :'</span>, t, <span class="kw">attributes</span>(t)<span class="op">$</span>units, <span class="st">'</span><span class="ch">\n</span><span class="st">'</span>)</code></pre></div>
-<pre><code>## time to complete : 12.99718 secs</code></pre>
+<pre><code>## time to complete : 16.61397 secs</code></pre>
 <p><br></p>
 <table>
 <caption>hamming-Cluster-Medoids</caption>

src/Makevars

@@ -1,4 +1,4 @@
-PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS)
+PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS) -DARMA_64BIT_WORD
 PKG_LIBS = $(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS) $(SHLIB_OPENMP_CXXFLAGS)
 CXX_STD = CXX11
 PKG_CPPFLAGS =	-I../inst/include/

src/Makevars.win

@@ -1,4 +1,4 @@
-PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS)
+PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS) -DARMA_64BIT_WORD
 PKG_LIBS = $(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS) $(SHLIB_OPENMP_CXXFLAGS) -mthreads
 CXX_STD = CXX11
 PKG_CPPFLAGS =	-I../inst/include/

src/kmeans_miniBatchKmeans_GMM_Medoids.cpp

@@ -546,8 +546,10 @@ Rcpp::List evaluation_rcpp(arma::mat& data, arma::vec CLUSTER, bool silhouette =
   Rcpp::List in_cluster_dist = INTRA_CLUSTER_DISS(data, tmp_clust);
 
   if (!silhouette) {
+
+    arma::rowvec befout_CLUSTER = arma::conv_to< arma::rowvec >::from(CLUSTER);
 
-    return(Rcpp::List::create(Rcpp::Named("clusters") = arma::conv_to< arma::rowvec >::from(CLUSTER),
+    return(Rcpp::List::create(Rcpp::Named("clusters") = befout_CLUSTER,
 
                               Rcpp::Named("cluster_indices") = tmp_clust,                                    // use the data indices, otherwise difficult to match clusters with silhouette or dissimilarity coefficients
 
@@ -557,8 +559,10 @@ Rcpp::List evaluation_rcpp(arma::mat& data, arma::vec CLUSTER, bool silhouette =
   else {
 
     Rcpp::List silhouet_out = SILHOUETTE_metric(data, CLUSTER, tmp_clust, in_cluster_dist);
+
+    arma::rowvec befout_CLUSTER = arma::conv_to< arma::rowvec >::from(CLUSTER);
 
-    return(Rcpp::List::create(Rcpp::Named("clusters") = arma::conv_to< arma::rowvec >::from(CLUSTER),
+    return(Rcpp::List::create(Rcpp::Named("clusters") = befout_CLUSTER,
 
                               Rcpp::Named("cluster_indices") = tmp_clust,                                       // use the data indices, otherwise difficult to match clusters with silhouette or dissimilarity coefficients
 
@@ -949,12 +953,20 @@ Rcpp::List GMM_arma(arma::mat& data, int gaussian_comps, std::string dist_mode,
   double n = timer.toc();
 
   if (verbose) { Rcpp::Rcout << "\ntime to complete : " << n << "\n" << std::endl; }
+
+  arma::mat model_means = model.means.t();
+
+  arma::mat model_dcovs = model.dcovs.t();
+
+  arma::rowvec model_hefts = arma::conv_to< arma::rowvec >::from(model.hefts.t());
+
+  double model_avg_log_p = model.avg_log_p(data.t(), gaussian_comps - 1);
 
-  return Rcpp::List::create( Rcpp::Named("centroids") = model.means.t(), Rcpp::Named("covariance_matrices") = model.dcovs.t(),           // each row of the 'covariance_matrices' is a different covariance matrix, use diag() to build each square diagonal matrix
+  return Rcpp::List::create( Rcpp::Named("centroids") = model_means, Rcpp::Named("covariance_matrices") = model_dcovs,           // each row of the 'covariance_matrices' is a different covariance matrix, use diag() to build each square diagonal matrix
 
-                             Rcpp::Named("weights") = model.hefts.t(), Rcpp::Named("Log_likelihood_raw") = loglik,
+                             Rcpp::Named("weights") = model_hefts, Rcpp::Named("Log_likelihood_raw") = loglik,
 
-                             Rcpp::Named("avg_Log_likelihood_DATA") = model.avg_log_p(data.t(), gaussian_comps - 1) );
+                             Rcpp::Named("avg_Log_likelihood_DATA") = model_avg_log_p );
 }
 
 
@@ -2261,7 +2273,18 @@ Rcpp::List ClusterMedoids(arma::mat& data, int clusters, std::string method, dou
     }
   }
 
-  if (clusters > 1) { silh_lst = silhouette_matrix(data, end_indices_vec + 1, end_cost_vec, threads); }
+  arma::mat befout_silhouette_matrix;
+
+  arma::mat befout_clustering_stats;
+
+  if (clusters > 1) { 
+
+    silh_lst = silhouette_matrix(data, end_indices_vec + 1, end_cost_vec, threads);
+
+    befout_silhouette_matrix = Rcpp::as<arma::mat> (silh_lst[0]);
+
+    befout_clustering_stats = Rcpp::as<arma::mat> (silh_lst[1]);
+  }
 
   arma::mat fuz_out;
 
@@ -2287,11 +2310,13 @@ Rcpp::List ClusterMedoids(arma::mat& data, int clusters, std::string method, dou
     }
   }
 
-  return Rcpp::List::create(Rcpp::Named("medoids") = end_idxs, Rcpp::Named("cost") = arma::accu(end_cost_vec), Rcpp::Named("dissimilarity_matrix") = data,
+  double end_cost_vec_scalar = arma::accu(end_cost_vec);
+
+  return Rcpp::List::create(Rcpp::Named("medoids") = end_idxs, Rcpp::Named("cost") = end_cost_vec_scalar, Rcpp::Named("dissimilarity_matrix") = data,
 
-                            Rcpp::Named("clusters") = end_indices_vec, Rcpp::Named("end_cost_vec") = end_cost_vec, Rcpp::Named("silhouette_matrix") = silh_lst[0],
+                            Rcpp::Named("clusters") = end_indices_vec, Rcpp::Named("end_cost_vec") = end_cost_vec, Rcpp::Named("silhouette_matrix") = befout_silhouette_matrix,
 
-                                        Rcpp::Named("fuzzy_probs") = fuz_out, Rcpp::Named("clustering_stats") = silh_lst[1], Rcpp::Named("flag_dissim_mat") = flag_dissim_mat);
+                                        Rcpp::Named("fuzzy_probs") = fuz_out, Rcpp::Named("clustering_stats") = befout_clustering_stats, Rcpp::Named("flag_dissim_mat") = flag_dissim_mat);
 }
 
 
@@ -2550,14 +2575,20 @@ Rcpp::List ClaraMedoids(arma::mat& data, int clusters, std::string method, int s
 
     bst_sample_silh_mat = Rcpp::as<arma::mat> (bst_lst[5]);
   }
+
+  arma::rowvec clr_split_out_rowvec = arma::conv_to< arma::rowvec >::from(clr_split_out);
+
+  arma::mat fuz_and_stats_mt = Rcpp::as<arma::mat> (fuz_and_stats[1]);
+
+  fuz_st_mat = fuz_st_mat.t();
 
   return Rcpp::List::create(Rcpp::Named("medoids") = subs_meds, Rcpp::Named("bst_dissimilarity") = dism, Rcpp::Named("medoid_indices") = out_medoid,
 
-                                        Rcpp::Named("sample_indices") = arma::conv_to< arma::rowvec >::from(clr_split_out), Rcpp::Named("clusters") = hard_clust,
-
-                                        Rcpp::Named("bst_sample_silhouette_matrix") = bst_sample_silh_mat, Rcpp::Named("fuzzy_probs") = fuz_and_stats[1],
+                                        Rcpp::Named("sample_indices") = clr_split_out_rowvec, Rcpp::Named("clusters") = hard_clust,
 
-                                                                                                                                                     Rcpp::Named("clustering_stats") = fuz_st_mat.t(), Rcpp::Named("bst_sample_dissimilarity_matrix") = bst_sample_dissm_mat);
+                                        Rcpp::Named("bst_sample_silhouette_matrix") = bst_sample_silh_mat, Rcpp::Named("fuzzy_probs") = fuz_and_stats_mt,
+
+                                        Rcpp::Named("clustering_stats") = fuz_st_mat, Rcpp::Named("bst_sample_dissimilarity_matrix") = bst_sample_dissm_mat);
 }
 
 
@@ -2645,18 +2676,22 @@ Rcpp::List split_rcpp_lst(Rcpp::List lst) {
   }
 
   double avg_intr_dis = arma::accu(intra_clust_disml);
+
+  double intra_clust_disml_scalar = avg_intr_dis;            // return this value
 
-  double avg_width_silh = arma::accu(silhouet);
+  double avg_width_silh = arma::accu(silhouet);    
 
-  avg_intr_dis /= silh_mat.n_rows;
+  avg_intr_dis /= silh_mat.n_rows;                           // modify this value
 
   avg_width_silh /= silh_mat.n_rows;
+
+  bool silh_plot_boolean = true;
 
   return Rcpp::List::create(Rcpp::Named("avg_intra_clust_dissimilarity") = avg_intr_dis,
 
-                            Rcpp::Named("sum_intra_dissim") = arma::accu(intra_clust_disml), Rcpp::Named("avg_width_silhouette") = avg_width_silh,
+                            Rcpp::Named("sum_intra_dissim") = intra_clust_disml_scalar, Rcpp::Named("avg_width_silhouette") = avg_width_silh,
 
-                            Rcpp::Named("list_intra_dissm") = intra_dissm, Rcpp::Named("list_silhouette") = silhouet_lst, Rcpp::Named("silhouette_plot") = true);
+                            Rcpp::Named("list_intra_dissm") = intra_dissm, Rcpp::Named("list_silhouette") = silhouet_lst, Rcpp::Named("silhouette_plot") = silh_plot_boolean);
 }