GMM Output - Teradata Vantage

Machine Learning Engine Analytic Function Reference
Product
Teradata Vantage
Release Number
8.10
1.1
Published
October 2019
Language
English (United States)
Last Update
2019-12-31
dita:mapPath
ima1540829771750.ditamap
dita:ditavalPath
jsj1481748799576.ditaval
dita:id
B700-4003
lifecycle
previous
Product Category
Teradata Vantage™

The GMM function outputs a message and a table. The OutputTable schema depends on the PackOutput syntax element.

Output Message Properties

Property Value
Output Table Table name specified in OutputTable syntax element.
Algorithm Used Algorithm that function used—Basic GMM or DP-DMM.
   
Stopping Criterion Why function stopped—reached iteration limit or convergence.
Delta Log Likelihood Change in mean log-likelihood for each data point between next-to-final and final iterations.
Number of Iterations Number of iterations that function performed before stopping.
Number of Clusters Found Number of clusters in GMM.
Covariance Type Spherical, diagonal, tied, or full.
   
Number of Data Points Number of data points in data set.
Global Mean Mean of data set.
Global Covariance Covariance of data set.
   
Log Likelihood Log-likelihood of data, given GMM.
Akaike Information Criterion Akaike Information Criterion.
Bayesian Information Criterion Bayesian Information Criterion.

OutputTable Schema, PackOutput ('false') (Default)

Column Data Type Description
cluster_id INTEGER Cluster identification number.
points_assigned INTEGER Number of points in training data set assigned to cluster.
covariance_type VARCHAR Covariance type.
weight DOUBLE PRECISION Weight assigned to cluster.
dim_n DOUBLE PRECISION [Column appears once for each dimension.] Mean of cluster n.
cov_n DOUBLE PRECISION [Column appears once for each dimension.] Covariance of cluster n.

Depends on CovarianceType:

CovarianceType cov_n
'spherical' Single covariance column. Each row contains DOUBLE PRECISION value.
'diagonal' D covariance rows, each containing a DOUBLE PRECISION value.
'tied' or 'full' D(D-1) covariance rows, each containing a DOUBLE PRECISION value.
log_determinant DOUBLE PRECISION Log determinant of covariance matrix.
prec VARCHAR Precision matrix, inverse of covariance matrix. Precision matrix is serialized and stored to improve performance of GMMPredict (ML Engine) function.

OutputTable Schema, PackOutput ('true')

Column Data Type Description
cluster_id INTEGER Cluster identification number.
points_assigned INTEGER Number of points in training data set assigned to cluster.
covariance_type VARCHAR Number of points in training data set assigned to cluster.
weight DOUBLE PRECISION Weight assigned to cluster.
mean VARCHAR Vector of D DOUBLE PRECISION values that specify mean of each cluster (for example, [4.5, 2.3, 1.3]).
covariance VARCHAR Depends on CovarianceType:
CovarianceType covariance
'spherical' Single covariance column. Each row contains DOUBLE PRECISION value.
'diagonal', 'tied', or 'full' Single covariance column. Each row contains a white-space separated list of D*D DOUBLE PRECISION values.
log_determinant DOUBLE PRECISION Log determinant of covariance matrix.
prec VARCHAR Precision matrix, inverse of covariance matrix. Precision matrix is serialized and stored to improve performance of GMMPredict (ML Engine) function.