Package smile.feature.extraction

Class BinaryEncoder

java.lang.Object

smile.feature.extraction.BinaryEncoder

All Implemented Interfaces:

Function<Tuple,int[]>

public class BinaryEncoder
extends Object
implements Function<Tuple,int[]>

Encodes categorical features using sparse one-hot scheme. The categorical attributes will be converted to binary dummy variables in a compact representation in which only indices of nonzero elements are stored in an integer array. In Maximum Entropy Classifier, the data are expected to store in this format.

Constructor Summary

Constructors

Constructor	Description
BinaryEncoder(StructType schema, String... columns)	Constructor.

Method Summary

Modifier and Type	Method	Description
int[][]	apply(DataFrame data)	Generates the compact representation of sparse binary features for a data frame.
int[]	apply(Tuple x)	Generates the compact representation of sparse binary features for given object.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface java.util.function.Function

andThen, compose

Constructor Details

BinaryEncoder

public BinaryEncoder(StructType schema,
 String... columns)

Constructor.

Parameters:

schema - the data frame schema.

columns - the column names of categorical variables. If empty, all categorical columns will be used.

Method Details

apply

public int[] apply(Tuple x)

Generates the compact representation of sparse binary features for given object.

Specified by:

apply in interface Function<Tuple,int[]>

Parameters:

x - an object of interest.

Returns:

an integer array of nonzero binary features.

apply

public int[][] apply(DataFrame data)

Generates the compact representation of sparse binary features for a data frame.

Parameters:

data - a data frame.

Returns:

the binary feature vectors.