The following node is available in the Open Source KNIME predictive analytics and data mining platform version 2.7.1. Discover over 1000 other nodes, as well as enterprise functionality at http://knime.com.

RProp MLP Learner

Implementation of the RProp algorithm for multilayer feedforward networks. RPROP performs a local adaptation of the weight-updates according to the behavior of the error function. For further details see: Riedmiller, M. Braun, H. : "A direct adaptive method for faster backpropagation learning: theRPROP algorithm",Proceedings of the IEEE International Conference on Neural Networks (ICNN) (Vol. 16, pp. 586-591). Piscataway, NJ: IEEE. This node provides a view of the error plot.
If the optional PMML inport is connected and contains preprocessing operations in the TransformationDictionary those are added to the learned model.

Dialog Options

Maximum number of iterations
The number of learning iterations.
Number of hidden layers
Specifies the number of hidden layers in the architecture of the neural network.
Number of hidden neurons per layer
Specifies the number of neurons contained in each hidden layer.
Class column
Choose the column that contains the target variable: it can either be nominal or numerical. All nominal class values are extracted and assigned to output neurons. If you use a numerical target variable (regression), please make sure it is normalized!
Ignore missing values
If this checkbox is set, rows with missing values will not be used for training.

Ports

Input Ports
0 Datatable with training data
1 Optional PMML port object containing preprocessing operations.
Output Ports
0 RProp trained Neural Network

Views

Error Plot
Displays the error for each iteration.
This node is contained in KNIME Base Nodes provided by KNIME GmbH, Konstanz, Germany.