System Administrator Guide > System Administrator Guide > Configuring Machine Learning Models (MLM) > Predefined UDFs for MLM > Predefined UDFs for ML Data Preprocessing
Was this helpful?
Predefined UDFs for ML Data Preprocessing
The following are the predefined UDFs for data preprocessing:
ACTIAN_BINARIZE
ACTIAN_BINARIZE(val, threshold)
Transforms the data into binary values or sets the feature value to 0 or 1 according to a threshold.
ACTIAN_LABEL_ENCODE
ACTIAN_LABEL_ENCODE(val)
Encodes labels of a column with values between 0 and n_classes-1. Labels may change between function runs if there is a change in the table.
ACTIAN_MAX_ABS_SCALE
ACTIAN_MAX_ABS_SCALE(val)
Scales values of a column by their maximum absolute value.
ACTIAN_MINMAX_SCALE
ACTIAN_MINMAX_SCALE(val)
Scales the values of a column based on the column's global minimum and maximum value.
ACTIAN_ONE_HOT_ENCODE
ACTIAN_ONE_HOT_ENCODE(val)
Encodes categorical features as a one-hot numeric array. For example,
AnalyticsEngine(actian_one_hot_encode(val), <distinct_count>, INT4)
Where, ‘distinct_count’ is a literal number determined by running the SQL query:
SELECT COUNT(DISTINCT val)
ACTIAN_ROBUST_SCALE
ACTIAN_ROBUST_SCALE(val)
Scales values of a column using statistics that are robust to outliers.
ACTIAN_STANDARD_SCALE_POP
ACTIAN_STANDARD_SCALE_POP(val)
Standardizes values of a column by removing the mean and scaling to unit variance using the population form of the standard deviation.
ACTIAN_STANDARD_SCALE_SAMP
ACTIAN_STANDARD_SCALE_SAMP(val)
Standardizes values of a column by removing the mean and scaling to unit variance using the sample form of the standard deviation.
ACTIAN_UNIVERSAL_SENTENCE_ENCODER
ACTIAN_UNIVERSAL_SENTENCE_ENCODER(input)
Embeds text into high-dimensional vectors that can be used for text classification, semantic similarity, clustering and other natural language understanding tasks. It is optimized for greater-than-word length text (such as sentences, phrases or short paragraphs) and data sources.
input is a column of the type string; output is a column of 512 dimensional vectors of the type float.
Note:  Function names may vary if, during an import, the user sets a function prefix that differs from the default.
Last modified date: 09/11/2026