Gaussian Process Classification Learner (sklearn)

Learns Gaussian Process Classification based on Laplace approximation implemented by scikit-learn library.

The implementation follows the algorithm in sections 3.1, 3.2 and 5.1 of the paper Gaussian Processes for Machine Learning by Carl E. Rasmussen and Christopher K.I. Williams (2006).

The model is trained with the selected nominal target column, and feature columns (can be nominal or numerical) from the input table. By default, the rightmost nominal column is selected as the target column and all the remaining nominal and numerical columns are selected as features.

Options

Input

Feature columns

Selection of columns used as feature columns. Columns with nominal and numerical data are allowed.

Target column

Selection of column used as the target column. Only columns with nominal data are allowed.

If there are missing values in the features

Define whether missing values in the input data should be skipped or whether the node execution should fail on missing values.

Available options:

  • Skip rows with missing values.: Rows with missing values will not be used for the training.
  • Fail on observing missing values.: Learner node will fail during the execution.

Algorithm Settings

Kernel

The kernel specifying the covariance function of the GP. The default kernel 1.0 * RBF(1.0) is used as default.

Available options:

  • Default: 1.0 * RBF(1.0)
  • Rational quadratic kernel: sklearn's RationalQuadratic kernel
  • Dot-Product Kernel: sklearn's DotProduct kernel
  • RBF: sklearn's Radial basis function kernel
  • White Kernel: sklearn's WhiteKernel
Multi-class classification method

Multi-class classification method selection.

Available options:

  • One-vs-Rest: One binary Gaussian process classifier is fitted for each class. This is the default.
  • One-vs-One: One binary Gaussian process classifier is fitted for each pair of classes

Input Ports

Icon

Nominal and numerical columns can be selected as feature columns from this table, and the target column must be nominal.

Output Ports

Icon

Trained Gaussian process classification model.

Popular Predecessors

  • No recommendations found

Popular Successors

  • No recommendations found

Views

This node has no views

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.