neuralnode.java

一个数据挖掘软件ALPHAMINERR的整个过程的JAVA版源代码

/*
 *    This program is free software; you can redistribute it and/or modify
 *    it under the terms of the GNU General Public License as published by
 *    the Free Software Foundation; either version 2 of the License, or
 *    (at your option) any later version.
 *
 *    This program is distributed in the hope that it will be useful,
 *    but WITHOUT ANY WARRANTY; without even the implied warranty of
 *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *    GNU General Public License for more details.
 *
 *    You should have received a copy of the GNU General Public License
 *    along with this program; if not, write to the Free Software
 *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/**
 * Title: XELOPES Data Mining Library
 * Description: The XELOPES library is an open platform-independent and data-source-independent library for Embedded Data Mining.
 * Copyright: Copyright (c) 2002 Prudential Systems Software GmbH
 * Company: ZSoft (www.zsoft.ru), Prudsys (www.prudsys.com)
 * @author Michael Thess
 * @version 1.2
 */
package com.prudsys.pdm.Models.Regression.NeuralNetwork;

import com.prudsys.pdm.Core.Category;
import com.prudsys.pdm.Core.MiningException;
import com.prudsys.pdm.Core.MiningHierarchyNode;
import com.prudsys.pdm.Core.PmmlPresentable;

/**
 * Base class for all neural nodes (neural input, neuron, neural output). <p>
 *
 * Mathematically the graph of a feedword neural network is a Directed
 * Acyclic Graph (DAG). Therefore this class extends MiningHierarchyNode.
 * We define: The inputs are the parents and the outputs are the children.
 * Thus the nodes of the neural input layer are the roots and the nodes of the
 * neural output layer are the leafs. The levels of the DAG correspond
 * to the layers of the neural network.
 */
public abstract class NeuralNode extends MiningHierarchyNode implements PmmlPresentable
{
  // -----------------------------------------------------------------------
  //  Constants of neural node connection type
  // -----------------------------------------------------------------------
  /** Disconnected type. */
  public static final int DISCONNECTED = 0;

  /** Connected type. */
  public static final int CONNECTED = 1;

  // -----------------------------------------------------------------------
  //  Variables declarations
  // -----------------------------------------------------------------------
  /** ID of the neural node. Is a unique identifier within the NN. */
  protected String id;

  /** Connection type. */
  protected int connectType = DISCONNECTED;

  /** Reference to NeuralLayer that owns this neural node. */
  protected NeuralLayer neuralLayer = null;

  /** IDs of input neurons. For PMML purposes. */
  protected String[] inputIDs = null;

  /** Input value of this node. */
  protected double input = Category.MISSING_VALUE;

  /** Output value of this node. */
  protected double output = Category.MISSING_VALUE;

  /** Error value of this node. */
  protected double error = Category.MISSING_VALUE;

  /** Have the weights of this neural unit already been updated? */
  protected boolean weightsUpdated = false;

  // -----------------------------------------------------------------------
  //  Constructor
  // -----------------------------------------------------------------------
  public NeuralNode() {
  }

  // -----------------------------------------------------------------------
  //  Getter and setter methods
  // -----------------------------------------------------------------------
  /**
   * Returns ID of neural node.
   *
   * @return ID of neural node
   */
  public String getId() {
    return id;
  }

  /**
   * Sets unique ID of neural node.
   *
   * @param id ID of neural node unique in Neural Network
   */
  public void setId(String id) {
    this.id = id;
  }

  /**
   * Returns connection type (disconnected, connected).
   *
   * @return connection type
   */
  public int getConnectType() {
    return connectType;
  }

  /**
   * Sets connection type (disconnected, connected).
   *
   * @param connectType new connection type
   */
  public void setConnectType(int connectType) {
    this.connectType = connectType;
  }

  /**
   * Returns neural layer object that owns this neural node.
   *
   * @return neural layer owner
   */
  public NeuralLayer getNeuralLayer() {
    return neuralLayer;
  }

  /**
   * Sets neural layer object that owns this neural node.
   *
   * @param neuralLayer new neural layer owner
   */
  public void setNeuralLayer(NeuralLayer neuralLayer) {
    this.neuralLayer = neuralLayer;
  }

  /**
   * Returns IDs of input neurons. Required for PMML.
   *
   * @return IDs of input neurons
   */
  public String[] getInputIDs() {
    return inputIDs;
  }

  // -----------------------------------------------------------------------
  //  Calculation methods
  // -----------------------------------------------------------------------
  /**
   * Reset all internal calculation values.
   */
  public abstract void resetValues();

  /**
   * Calculates net input of this node.
   *
   * @param newcalc calculate the input new
   * @return input value, missing if not calculated
   * @exception MiningException error while calculating input value
   */
  public abstract double inputValue(boolean newcalc) throws MiningException;

  /**
   * Calculates output of this node. This is the activation function
   * applied to the input value.
   *
   * @param newcalc calculate the output new
   * @return output value, missing if not calculated
   * @exception MiningException error while calculating output value
   */
  public abstract double outputValue(boolean newcalc) throws MiningException;

  /**
   * Calculates error of this node.
   *
   * @param newcalc calculate the error new
   * @return error value, missing if not calculated
   * @exception MiningException error while calculating error value
   */
  public abstract double errorValue(boolean newcalc) throws MiningException;

  /**
   * Returns weight at specified index. Should be overriden by Neuron's.
   *
   * @param index weight index
   * @return weight, missing if invalid index
   */
  double getWeightAt(int index) {
      return 1.0;
  }

  /**
   * This method updates the weights of this neural node and then
   * calls the same method of the input nodes recursively. Should
   * be overriden by Neuron's.
   *
   * @param learningRate the learning rate
   * @param momentum the momentum
   * @exception MiningException cannot update weights
   */
  public void updateWeights(double learningRate, double momentum)
    throws MiningException {

    if (!weightsUpdated) {
      for (int i = 0; i < getParentsCount(); i++)
        ((NeuralNode) getParentAt(i)).updateWeights(learningRate, momentum);

      weightsUpdated = true;
    }
  }

  // -----------------------------------------------------------------------
  //  Further methods
  // -----------------------------------------------------------------------
  /**
   * Delivers string representation of neural node.
   *
   * @return string representation of neural node
   */
  public String toString() {

    String s = "ID = " + id;
    int npars = getParentsCount();
    if (npars > 0) {
      s = s + " inputs: ";
      for (int i = 0; i < npars; i++)
        s = s + ((NeuralNode) getParentAt(i)).getId() + " ";
    };

    return s;
  }

}