Arguments - Aster Analytics

Teradata Aster Analytics Foundation User Guide
Product
Aster Analytics
Release Number
6.21
Published
November 2016
Language
English (United States)
Last Update
2018-04-14
dita:mapPath
kiu1466024880662.ditamap
dita:ditavalPath
AA-notempfilter_pdf_output.ditaval
dita:id
B700-1021
lifecycle
previous
Product Category
Software
Argument Category Description
InputTable Required Input table is the table containing the list of features by which to cluster the data.
OutputTable Required Output table is the table where output is stored. The output table contains the centroids of the clusters.
InitialSeedTable Required if NumClusters is omitted, otherwise not allowed An input table containing the points that serve as initial cluster centers.
NumClusters Required if InitialSeedTable is omitted, otherwise not allowed If a single value is given, the function trains a model with that number of clusters. If a list of integers is supplied, the function trains a model for each value.

Initial seeds are specified by performing KMeans|| sampling using the FixedSample function.
ModelIdColumn Optional If this argument is present, it indicates that the table specified in InitialSeedTable contains more than one set of seed values (that is, it contains seed values for more than one model).

This argument specifies the column in InitialSeedTable that identifies which rows are associated with each model.
InputColumns Required Specifies the input table columns to use for clustering.
Threshold Optional This is the convergence threshold. When the centroids move by less than this amount, the algorithm has converged. The input value must be no less than 0.0. The default value is 0.0395.
MaxIterNum Optional Specifies the maximum number of iterations that the algorithm runs before quitting if the convergence threshold is not met. The input value must be an integer greater than 0. The default value is 10.
Distance Optional Specifies the distance metric that the Kmodes function uses for numeric dimensions. The default value is 'euclidean'.
CategoricalDistance Optional Specifies the distance metric that the Kmodes function uses for categorical dimensions:
  • 'overlap' (default):

    Distance is 0 if two points are in the same category or 1 if they are in different categories.

  • 'hamming':

    Used for categories that are strings of equal length. The fraction of characters that are different.
CategoryWeights Optional The weights to be assigned to each category in the KModes distance.
AsCategories Optional Indicates which numeric categories to interpret as categorical variables. Input columns must contain numeric SQL types.