smile.data

Class Datum<T>

java.lang.Object

smile.data.Datum<T>

public class Datum<T>
extends Object

An object of generic datum and optional weight. Although the data object is immutable itself, the associated properties such as class label/response weight are mutable.

Author:

Haifeng Li

Field Summary

Fields

Modifier and Type	Field and Description
String	description Optional detailed description.
String	name Name of datum.
Timestamp	timestamp Timestamp of datum in case of transactional data.
double	weight Optional weight of this datum.
T	x Immutable datum object.
double	y Class label or real-valued response.

Constructor Summary

Constructors

Constructor and Description
Datum(T x) Constructor.
Datum(T x, double y) Constructor.
Datum(T x, double y, double weight) Constructor.

Method Summary

Methods inherited from class java.lang.Object

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

Field Detail

x

public final T x

Immutable datum object.

y

public double y

Class label or real-valued response. NaN means unknown label/response.

weight

public double weight

Optional weight of this datum. By default, it is 1.0. The particular meaning of weight depends on applications and machine learning algorithms. Although there are on explicit requirements on the weights, in general, they should be positive.

name

public String name

Name of datum.

description

public String description

Optional detailed description.

timestamp

public Timestamp timestamp

Timestamp of datum in case of transactional data. Transactional data are time-stamped data collected over time at no particular frequency. Some examples of transactional data are

• Internet data
• Point of Sales (POS) data
• Inventory data
• Call Center data
• Trading data

Constructor Detail

Datum

public Datum(T x)

Constructor.

Parameters:

x - the datum.

Datum

public Datum(T x,
     double y)

Constructor.

Parameters:

x - the datum.

y - the class label or real-valued response.

Datum

public Datum(T x,
     double y,
     double weight)

Constructor.

Parameters:

x - the datum.

y - the class label or real-valued response.

weight - the weight of datum. The particular meaning of weight depends on applications and machine learning algorithms. Although there are on explicit requirements on the weights, in general, they should be positive.