com.github.waikatodatamining.matrix.algorithm.AbstractMultiReponsePLS Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of matrix-algorithms Show documentation
Java library of 2-dimensional matrix algorithms.
There is a newer version: 0.1.0
/*
 *   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 3 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, see .
 */

/*
 * AbstractSingleReponsePLS.java
 * Copyright (C) 2018 University of Waikato, Hamilton, NZ
 */

package com.github.waikatodatamining.matrix.algorithm;

import Jama.Matrix;
import com.github.waikatodatamining.matrix.core.MatrixHelper;
import com.github.waikatodatamining.matrix.transformation.AbstractTransformation;
import com.github.waikatodatamining.matrix.transformation.Center;
import com.github.waikatodatamining.matrix.transformation.Standardize;

/**
 * Ancestor for PLS algorithms that work on a single response variable.
 *
 * @author FracPete (fracpete at waikato dot ac dot nz)
 */
public abstract class AbstractMultiReponsePLS
  extends AbstractPLS {

  private static final long serialVersionUID = -8160023117935320371L;

  /** the class mean. */
  protected double[] m_ClassMean;

  /** the class stddev. */
  protected double[] m_ClassStdDev;

  /** the transformation for the predictors. */
  protected AbstractTransformation m_TransPredictors;

  /** the transformation for the response. */
  protected AbstractTransformation m_TransResponse;

  /**
   * Resets the member variables.
   */
  @Override
  protected void reset() {
    super.reset();

    m_ClassMean       = null;
    m_ClassStdDev     = null;
    m_TransPredictors = null;
    m_TransResponse   = null;
  }

  /**
   * Returns the minimum number of columns the response matrix has to have.
   *
   * @return		the minimum
   */
  protected abstract int getMinColumnsResponse();

  /**
   * Returns the maximum number of columns the response matrix has to have.
   *
   * @return		the maximum, -1 for unlimited
   */
  protected abstract int getMaxColumnsResponse();

  /**
   * Hook method for checking the data before training.
   *
   * @param predictors	the input data
   * @param response 	the dependent variable(s)
   * @return		null if successful, otherwise error message
   */
  @Override
  protected String check(Matrix predictors, Matrix response) {
    String	result;

    result = super.check(predictors, response);

    if (result == null) {
      if (response.getColumnDimension() < getMinColumnsResponse())
	result = "Algorithm requires at least " + getMinColumnsResponse() + " response columns, found: " + response.getColumnDimension();
      else if ((getMaxColumnsResponse() != -1) && (response.getColumnDimension() > getMaxColumnsResponse()))
	result = "Algorithm can handle at most " + getMaxColumnsResponse() + " response columns, found: " + response.getColumnDimension();
    }

    return result;
  }

  /**
   * Initializes using the provided data.
   *
   * @param predictors	the input data
   * @param response 	the dependent variable(s)
   * @return		null if successful, otherwise error message
   * @throws Exception	if analysis fails
   */
  protected abstract String doPerformInitialization(Matrix predictors, Matrix response) throws Exception;

  /**
   * Initializes using the provided data.
   *
   * @param predictors	the input data
   * @param response 	the dependent variable(s)
   * @return		null if successful, otherwise error message
   * @throws Exception	if analysis fails
   */
  protected String doInitialize(Matrix predictors, Matrix response) throws Exception {
    String	result;
    int		i;

    m_ClassMean   = new double[response.getColumnDimension()];
    m_ClassStdDev = new double[response.getColumnDimension()];
    for (i = 0; i < response.getColumnDimension(); i++) {
      switch (m_PreprocessingType) {
	case CENTER:
	  m_ClassMean[i]   = MatrixHelper.mean(response, 0);
	  m_ClassStdDev[i] = 1;
	  m_TransPredictors = new Center();
	  m_TransResponse   = new Center();
	  break;
	case STANDARDIZE:
	  m_ClassMean[i]    = MatrixHelper.mean(response, 0);
	  m_ClassStdDev[i]  = MatrixHelper.stdev(response, 0);
	  m_TransPredictors = new Standardize();
	  m_TransResponse   = new Standardize();
	  break;
	case NONE:
	  m_ClassMean[i]    = 0;
	  m_ClassStdDev[i]  = 1;
	  m_TransPredictors = null;
	  m_TransResponse   = null;
	  break;
	default:
	  throw new IllegalStateException("Unhandled preprocessing type; " + m_PreprocessingType);
      }
    }

    if (m_TransPredictors != null) {
      m_TransPredictors.configure(predictors);
      predictors = m_TransPredictors.transform(predictors);
    }
    if (m_TransResponse != null) {
      m_TransResponse.configure(response);
      response = m_TransResponse.transform(response);
    }

    result = doPerformInitialization(predictors, response);

    return result;
  }

  /**
   * Performs predictions on the data.
   *
   * @param predictors the input data
   * @throws Exception if analysis fails
   * @return the transformed data and the predictions
   */
  protected abstract Matrix doPerformPredictions(Matrix predictors) throws Exception;

  /**
   * Performs predictions on the data.
   *
   * @param predictors the input data
   * @throws Exception if analysis fails
   * @return the transformed data and the predictions
   */
  @Override
  protected Matrix doPredict(Matrix predictors) throws Exception {
    Matrix	result;
    int		i;
    int		j;

    result = doPerformPredictions(predictors);
    if (m_TransResponse != null) {
      for (i = 0; i < result.getRowDimension(); i++) {
	for (j = 0; j < result.getColumnDimension(); j++)
	  result.set(i, j, result.get(i, j) * m_ClassStdDev[j] + m_ClassMean[j]);
      }
    }

    return result;
  }
}