/*
 * Copyright (C) 2008-2015 by Holger Arndt
 *
 * This file is part of the Java Data Mining Package (JDMP).
 * See the NOTICE file distributed with this work for additional
 * information regarding copyright ownership and licensing.
 *
 * JDMP is free software; you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License as
 * published by the Free Software Foundation; either version 2
 * of the License, or (at your option) any later version.
 *
 * JDMP 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 Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with JDMP; if not, write to the
 * Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
 * Boston, MA  02110-1301  USA
 */

package org.jdmp.core.algorithm.regression;

import org.jdmp.core.dataset.ListDataSet;
import org.jdmp.core.sample.Sample;
import org.jdmp.core.variable.Variable;
import org.ujmp.core.Matrix;
import org.ujmp.core.calculation.Calculation.Ret;

/**
 * AlgorithmLinearRegression extends AlgorithmClassifier and not
 * AlgorithmRegression because also classification is possible using regression
 */
public class LinearRegression extends AbstractRegressor {
	private static final long serialVersionUID = 3483912497269476834L;

	public static final String PARAMETERS = "Parameters";

	private Matrix mean = null;
	private Matrix std = null;

	private final int numberOfPrincipalComponents;

	public LinearRegression() {
		this(0);
	}

	public LinearRegression(int numberOfPrincipalComponents) {
		super();
		this.numberOfPrincipalComponents = numberOfPrincipalComponents;
		setParameterVariable(Variable.Factory.labeledVariable("Regression Parameters"));
	}

	public void setParameterVariable(Variable variable) {
		setVariable(PARAMETERS, variable);
	}

	public Variable getParameterVariable() {
		return getVariableMap().get(PARAMETERS);
	}

	public Matrix getParameterMatrix() {
		return getParameterVariable().getLast();
	}

	public Matrix predictOne(Matrix input) {
		input = input.toColumnVector(Ret.NEW);

		Matrix x = Matrix.Factory.zeros(1, input.getColumnCount() + 1);
		for (int c = 0; c < input.getColumnCount(); c++) {
			x.setAsDouble(input.getAsDouble(0, c), 0, c + 1);
		}

		x = x.minus(mean).divide(std);
		x.setAsDouble(1, 0, 0);

		Matrix result = x.mtimes(getParameterMatrix());
		return result;
	}

	public void trainAll(ListDataSet dataSet) {
		System.out.println("training started");

		final int featureCount = getFeatureCount(dataSet);
		final int classCount = getClassCount(dataSet);
		final int sampleCount = dataSet.size();

		Matrix x = Matrix.Factory.zeros(dataSet.size(), featureCount + 1);
		Matrix y = Matrix.Factory.zeros(dataSet.size(), classCount);

		int i = 0;
		for (Sample s : dataSet) {
			Matrix input = s.getAsMatrix(getInputLabel()).toColumnVector(Ret.LINK);
			for (int c = 0; c < input.getColumnCount(); c++) {
				x.setAsDouble(input.getAsDouble(0, c), i, c + 1);
			}
			Matrix target = s.getAsMatrix(getTargetLabel()).toColumnVector(Ret.LINK);
			for (int c = 0; c < target.getColumnCount(); c++) {
				y.setAsDouble(target.getAsDouble(0, c), i, c);
			}
			i++;
		}

		mean = x.mean(Ret.NEW, ROW, true);

		for (int r = 0; r < x.getRowCount(); r++) {
			for (int c = 0; c < x.getColumnCount(); c++) {
				x.setAsDouble(x.getAsDouble(r, c) - mean.getAsDouble(0, c), r, c);
			}
		}

		std = x.std(Ret.NEW, ROW, true, true);

		for (int r = 0; r < x.getRowCount(); r++) {
			for (int c = 0; c < x.getColumnCount(); c++) {
				double s = std.getAsDouble(0, c);
				if (s == 0) {
					x.setAsDouble(x.getAsDouble(r, c), r, c);
				} else {
					x.setAsDouble(x.getAsDouble(r, c) / s, r, c);
				}
			}
		}

		for (int r = 0; r < x.getRowCount(); r++) {
			x.setAsDouble(1, r, 0);
		}

		System.out.println("data loaded");

		Matrix parameters;

		if (numberOfPrincipalComponents > 0
				&& numberOfPrincipalComponents < Math.min(x.getRowCount(), x.getColumnCount() - 1)) {
			if (sampleCount < featureCount) {
				// final Matrix xt = x.transpose();
				parameters = x.pinv(numberOfPrincipalComponents).mtimes(y);
			} else {
				final Matrix xt = x.transpose();
				parameters = xt.mtimes(x).pinv(numberOfPrincipalComponents).mtimes(xt).mtimes(y);
			}
		} else {
			if (sampleCount < featureCount) {
				parameters = x.pinv().mtimes(y);
			} else {
				final Matrix xt = x.transpose();
				parameters = xt.mtimes(x).pinv().mtimes(xt).mtimes(y);
			}
		}

		System.out.println("training finished");

		setParameterMatrix(parameters);
	}

	public void setParameterMatrix(Matrix parameters) {
		getVariableMap().setMatrix(PARAMETERS, parameters);
	}

	public void reset() {
		getParameterVariable().clear();
	}

	public Regressor emptyCopy() {
		LinearRegression lr = new LinearRegression(numberOfPrincipalComponents);
		lr.setInputLabel(getInputLabel());
		lr.setTargetLabel(getTargetLabel());
		return lr;
	}

}