Vector Example Using TD_KMeans - Analytics Database

InputTable

DROP TABLE input_table_udt;
CREATE MULTISET TABLE input_table_udt(id INTEGER, array_col Vector) PRIMARY
INDEX ( id );
INSERT INTO input_table_udt values(1, '-0.06944,0.080352,0.045963,0.006985,-0.000496');
INSERT INTO input_table_udt values(2, '-0.025461,0.045302,-0.028107,0.031248,0.00077');
INSERT INTO input_table_udt values(3, '-0.028107,0.016781,-0.023147,-0.068779,-0.07936');
INSERT INTO input_table_udt values(4, '-0.030091,0.027611,-0.047616,-0.025461,0.029595');

IntialCentroidsTable

DROP TABLE initial_centroids_table_udt;
CREATE MULTISET TABLE initial_centroids_table_udt as (SELECT * FROM
input_table_udt where id in (2,4)) with data;

Example: TD_KMeans SQL Call: InitialCentroidsTable not provided

SELECT td_clusterid_kmeans, cast(array_col as Varchar(100)) as array_col, td_size_kmeans, td_withinss_kmeans, id, td_modelinfo_kmeans
FROM TD_KMeans (
ON input_table_udt as InputTable
USING
IdColumn('id')
TargetColumns('array_col')
NumClusters(2)
Seed(0)
StopThreshold(0.0395)
MaxIterNum(3)
EmbeddingSize(5)
)as dt;

Output: InitialCentroidsTable not provided

Example: TD_KMeans SQL Call: InitialCentroidsTable is provided

SELECT td_clusterid_kmeans, cast(array_col as Varchar(100)) as array_col, td_size_kmeans, td_withinss_kmeans, id, td_modelinfo_kmeans FROM TD_KMeans (
ON input_table_udt as InputTable
ON initial_centroids_table_udt as InitialCentroidsTable DIMENSION
USING
IdColumn('id')
TargetColumns('array_col')
StopThreshold(0.0395)
MaxIterNum(3)
EmbeddingSize(5)
)as dt;

Output: InitialCentroidsTable is provided