3  Clustering

using Rtemis

Get available clutering algorithms:

clustselect()

3.1 Data

iris = ddb_data("~/icloud/Data/iris.csv")

Create a 2D projection for plotting of clustering results using a 2D scatter

iris_umap = d_UMAP(iris[:, 1:4], 2)

3.2 K-means

iris_kmeans = c_KMEANS(iris[:, 1:4], 3);

dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_kmeans.clusters,
    trace_names = ["K-means"])

3.3 K-medoids

iris_kmedoids = c_KMEDOIDS(iris[:, 1:4], 3);

dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_kmedoids.clusters,
    trace_names = ["K-medoids"])

3.4 Fuzzy C-Means

iris_cmeans = c_CMEANS(iris[:, 1:4], 3,
    fuzziness=14, trace=1);
dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_cmeans.clusters,
    trace_names = ["C-means"])

dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_cmeans.clusters,
    trace_names = ["C-means"])

3.5 DBSCAN

iris_dbscan = c_DBSCAN(iris[:, 1:4])

dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_dbscan.clusters,
    trace_names = ["DBSCAN"])

3.6 MCL

iris_mcl = c_MCL(iris[:, 1:4])

dplot3_xy(
    iris_umap.projections_train[:, 1], 
    iris_umap.projections_train[:, 2],
    group = iris_mcl.clusters,
    trace_names = ["MCL"],
    palette = repeat(Rtemis.rtcol3, 4))