Skip to content

[source]

Hyperplane#

Generates a labeled dataset whose samples form a hyperplane in n-dimensional vector space and whose labels are continuous values drawn from a uniform random distribution between -1 and 1. When the number of coefficients is either 1, 2 or 3, the samples form points, lines, and planes respectively. Due to its linearity, Hyperplane is especially useful for testing linear regression models.

Data Types: Continuous

Label Type: Continuous

Parameters#

# Name Default Type Description
1 coefficients [1, -1] array The n coefficients of the hyperplane where n is the dimensionality.
2 intercept 0.0 float The y intercept term.
3 noise 0.1 float The factor of gaussian noise to add to the data points.

Example#

use Rubix\ML\Datasets\Generators\Hyperplane;

$generator = new Hyperplane([0.1, 3, -5, 0.01], 150.0, 0.25);

Additional Methods#

This generator does not have any additional methods.


Last update: 2021-01-23