analysis/org.eclipse.stem.util.analysis/src/org/eclipse/stem/util/analysis/ParameterEstimator.java - stem/org.eclipse.stem - Git at Google


 package org.eclipse.stem.util.analysis;

 /*******************************************************************************
  * Copyright (c) 2007, 2008 IBM Corporation and others.
  * All rights reserved. This program and the accompanying materials
  * are made available under the terms of the Eclipse Public License v1.0
  * which accompanies this distribution, and is available at
  * http://www.eclipse.org/legal/epl-v10.html
  *
  * Contributors:
  *     IBM Corporation - initial API and implementation
  *******************************************************************************/


 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.List;
 import java.util.Map;

 import org.eclipse.stem.analysis.impl.ReferenceScenarioDataMapImpl.ReferenceScenarioDataInstance;
 import org.eclipse.stem.util.analysis.views.EstimatorControl;

 /**
  * Estimate parameters for an SI, SIR or SEIR model
  *
  */

 public abstract class ParameterEstimator {

 	protected static final String keyS = "S";
 	protected static final String keyE = "E";
 	protected static final String keyI = "I";
 	protected static final String keyR = "R";
 	protected static final String keySelectedS = "S*";
 	protected static final String keySelectedE = "E*";
 	protected static final String keySelectedI = "I*";
 	protected static final String keySelectedR = "R*";

 	protected static final String keyXY1 = "X1vsY1";
 	protected static final String keyXY2 = "X2vsY2";
 	protected static final String keyXY3 = "X3vsY3";
 	protected static final String keyXY4 = "X4vsY4";

 	protected static final String keyFit1 = "Fit XY1";
 	protected static final String keyFit2 = "Fit XY2";
 	protected static final String keyFit3 = "Fit XY3";
 	protected static final String keyFit4 = "Fit XY4";


 	protected static double minSthreshold = 0.01;
 	protected static double minEthreshold = 0.01;
 	protected static double minIthreshold = 0.01;
 	protected static double minRthreshold = 0.01;

 	protected static double maxSthreshold = 0.99;
 	protected static double maxEthreshold = 0.99;
 	protected static double maxIthreshold = 0.99;
 	protected static double maxRthreshold = 0.99;

 	protected static boolean useFreqDependantBeta = true;


 	/**
 	 * rejected data where fit failed
 	 */
 	protected double rejectCount = 0.0;

 	/*
 	 * the SELECTED data to plot
 	 */
 	List<double[][]>  dataToPlot = new ArrayList<double[][]>();


 	List<double[][]> regressionValues = new ArrayList<double[][]>();


 	List<double[][]> fittedValues = new ArrayList<double[][]>();


 	protected ReferenceScenarioDataInstance data;

 	/**
 	 * Sets the minimum number of points needed to do a reasonable fit
 	 * TODO should not be hard coded
 	 */
 	protected static final int MINSIZE = 7;

 	/**
 	 * If derivativeMethod = LINEAR_DERIVATIVE  then a linear algorithm
 	 * is used to compute derivatives
 	 */
 	public static final int LINEAR_DERIVATIVE = 1;


 	/**
 	 * If derivativeMethod = POLYNOMIAL_DERIVATIVE  then a polynomial algorithm
 	 * is used to compute derivatives
 	 */
 	public static final int POLYNOMIAL_DERIVATIVE = 2;

 	// All models have these vectors
 	double[] p;
 	double[] s;
 	double[] i;
 	double[] t;

 	/**
 	 * set to true only if there are enough points to do a reasonable fit
 	 */
 	public boolean dataValid = false;

 	int numDataPoints;


 	/**
 	 * construct a new estimator for the input data[] SIR
 	 * @param data
 	 */
 	public ParameterEstimator(ReferenceScenarioDataInstance data) {
 		this.data = data;
 		// Always Prepare the data for analysis
 		this.data.findMaxValues();
 	}

 	/**
 	 * The control needs this information but must get it through the analyzer from the Estimator
 	 * created by the factory
 	 * the number of differential equations in the solution
 	 * @return number of equations determines the number of graphs
 	 */
 	public abstract int getNumProperties();

 	/**
 	 * The control needs this information but must get it through the analyzer from the Estimator
 	 * created by the factory
 	 * @param equationIndex , one chart per state S, E, I, and R.
 	 * @param state 0 or 1 representing S and Sselected, I and I selected,etc.
 	 * @return property name
 	 */
 	public String getProperty(int equationIndex, int state) {
 		 switch (equationIndex) {
          case 0:
         	 if(state==0) {
         		 return keyS;
         	 } else if (state==1) {
         		 return keySelectedS;
         	 } else if ((state==2)||(state==3)) {
         		 return keyXY1;
         	 }else  {
         		 return keyFit1;
         	 }

          case 1:
         	 if(state==0) {
         		 return keyI;
         	 } else if (state==1) {
         		 return keySelectedI;
         	 } else if ((state==2)||(state==3)) {
         		 return keyXY2;
         	 }else  {
         		 return keyFit2;
         	 }

          case 2:
         	 if(state==0) {
         		 return keyR;
         	 } else if (state==1) {
         		 return keySelectedR;
         	 } else if ((state==2)||(state==3)) {
         		 return keyXY3;
         	 }else  {
         		 return keyFit3;
         	 }

          case 3:
         	 if(state==0) {
         		 return keyE;
         	 } else if (state==1) {
         		 return keySelectedE;
         	 } else if ((state==2)||(state==3)) {
         		 return keyXY4;
         	 }else  {
         		 return keyFit4;
         	 }

          default: return null;
 		 }// switch
 	}// getProperty

 	/**
 	 * update the min and max thresholds from the
 	 * EstimatorControl UI spinners
 	 * @param controlParametersMap
 	 */
 	public void setAllParameters(Map<String,Object> controlParametersMap) {
 		if(controlParametersMap  !=null) {
 			maxSthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_S_KEY)).doubleValue();
 			minSthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_S_KEY)).doubleValue();

 			maxEthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_E_KEY)).doubleValue();
 			minEthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_E_KEY)).doubleValue();

 			maxIthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_I_KEY)).doubleValue();
 			minIthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_I_KEY)).doubleValue();

 			maxRthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_R_KEY)).doubleValue();
 			minRthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_R_KEY)).doubleValue();

 			useFreqDependantBeta = ((Boolean) controlParametersMap.get(EstimatorControl.FREQ_MODE_KEY) ).booleanValue();

 		}
 		return;
 	}// update Thresholds


 	/**
 	 * Estimate the parameters
 	 * @param localDensity
 	 * @param localPopulation
 	 * @param locationID
 	 * @return the ModelParameters estimate
 	 */
 	public abstract ModelParameters estimate(double localDensity, double localPopulation, String locationID) ;


 	/**
 	 * Returns the raw data (state 0) and the selected data (state 1)
 	 * for each of the state variables S,E,I,R (chartIndex 0-3)
 	 * @param chartIndex
 	 * @param state
 	 * @return the data to plot
 	 */
 	public double[] getValues(int chartIndex, int state) {
 		// pick the right data for the right chart
 		if(state <= 1) {
 			double[][] data = dataToPlot.get(chartIndex);
 			return data[state];
 		}
 		// regression data
 		double[][] regression = regressionValues.get(chartIndex);
 		return regression[state-2];
 	}// getValues


 	/**
 	 * create a sorted regression (ordered by the x coordinate)
 	 * @param xList
 	 * @param yList
 	 * @return the regression
 	 */
 	public double[][] getSortedRegression(List<Double> xList, List<Double> yList) {
 		double[][] regression = new double[2][xList.size()];
 		if(yList.size() != xList.size()) return regression;
 		List<PhaseSpaceCoordinate> coords = new ArrayList<PhaseSpaceCoordinate>();
 		for (int i = 0; i < xList.size(); i ++) {
 			double x = xList.get(i).doubleValue();
 			double y = yList.get(i).doubleValue();
 			PhaseSpaceCoordinate coord = new PhaseSpaceCoordinate(x,y);
 			coords.add(coord);
 		}
 		Object[] coordArray = coords.toArray();
 		Arrays.sort(coordArray);
 		for (int i = 0; i < xList.size(); i ++) {
 			PhaseSpaceCoordinate coord = (PhaseSpaceCoordinate)coordArray[i];
 			regression[0][i] = coord.xValue;
 			regression[1][i] = coord.yValue;
 		}
 		return regression;
 	}//getSortedRegression

 }

	package org.eclipse.stem.util.analysis;

	/*******************************************************************************
	* Copyright (c) 2007, 2008 IBM Corporation and others.
	* All rights reserved. This program and the accompanying materials
	* are made available under the terms of the Eclipse Public License v1.0
	* which accompanies this distribution, and is available at
	* http://www.eclipse.org/legal/epl-v10.html
	*
	* Contributors:
	* IBM Corporation - initial API and implementation
	*******************************************************************************/


	import java.util.ArrayList;
	import java.util.Arrays;
	import java.util.List;
	import java.util.Map;

	import org.eclipse.stem.analysis.impl.ReferenceScenarioDataMapImpl.ReferenceScenarioDataInstance;
	import org.eclipse.stem.util.analysis.views.EstimatorControl;

	/**
	* Estimate parameters for an SI, SIR or SEIR model
	*
	*/

	public abstract class ParameterEstimator {

	protected static final String keyS = "S";
	protected static final String keyE = "E";
	protected static final String keyI = "I";
	protected static final String keyR = "R";
	protected static final String keySelectedS = "S*";
	protected static final String keySelectedE = "E*";
	protected static final String keySelectedI = "I*";
	protected static final String keySelectedR = "R*";

	protected static final String keyXY1 = "X1vsY1";
	protected static final String keyXY2 = "X2vsY2";
	protected static final String keyXY3 = "X3vsY3";
	protected static final String keyXY4 = "X4vsY4";

	protected static final String keyFit1 = "Fit XY1";
	protected static final String keyFit2 = "Fit XY2";
	protected static final String keyFit3 = "Fit XY3";
	protected static final String keyFit4 = "Fit XY4";


	protected static double minSthreshold = 0.01;
	protected static double minEthreshold = 0.01;
	protected static double minIthreshold = 0.01;
	protected static double minRthreshold = 0.01;

	protected static double maxSthreshold = 0.99;
	protected static double maxEthreshold = 0.99;
	protected static double maxIthreshold = 0.99;
	protected static double maxRthreshold = 0.99;

	protected static boolean useFreqDependantBeta = true;


	/**
	* rejected data where fit failed
	*/
	protected double rejectCount = 0.0;

	/*
	* the SELECTED data to plot
	*/
	List<double[][]> dataToPlot = new ArrayList<double[][]>();


	List<double[][]> regressionValues = new ArrayList<double[][]>();


	List<double[][]> fittedValues = new ArrayList<double[][]>();


	protected ReferenceScenarioDataInstance data;

	/**
	* Sets the minimum number of points needed to do a reasonable fit
	* TODO should not be hard coded
	*/
	protected static final int MINSIZE = 7;

	/**
	* If derivativeMethod = LINEAR_DERIVATIVE then a linear algorithm
	* is used to compute derivatives
	*/
	public static final int LINEAR_DERIVATIVE = 1;


	/**
	* If derivativeMethod = POLYNOMIAL_DERIVATIVE then a polynomial algorithm
	* is used to compute derivatives
	*/
	public static final int POLYNOMIAL_DERIVATIVE = 2;

	// All models have these vectors
	double[] p;
	double[] s;
	double[] i;
	double[] t;

	/**
	* set to true only if there are enough points to do a reasonable fit
	*/
	public boolean dataValid = false;

	int numDataPoints;


	/**
	* construct a new estimator for the input data[] SIR
	* @param data
	*/
	public ParameterEstimator(ReferenceScenarioDataInstance data) {
	this.data = data;
	// Always Prepare the data for analysis
	this.data.findMaxValues();
	}

	/**
	* The control needs this information but must get it through the analyzer from the Estimator
	* created by the factory
	* the number of differential equations in the solution
	* @return number of equations determines the number of graphs
	*/
	public abstract int getNumProperties();

	/**
	* The control needs this information but must get it through the analyzer from the Estimator
	* created by the factory
	* @param equationIndex , one chart per state S, E, I, and R.
	* @param state 0 or 1 representing S and Sselected, I and I selected,etc.
	* @return property name
	*/
	public String getProperty(int equationIndex, int state) {
	switch (equationIndex) {
	case 0:
	if(state==0) {
	return keyS;
	} else if (state==1) {
	return keySelectedS;
	} else if ((state==2)\|\|(state==3)) {
	return keyXY1;
	}else {
	return keyFit1;
	}

	case 1:
	if(state==0) {
	return keyI;
	} else if (state==1) {
	return keySelectedI;
	} else if ((state==2)\|\|(state==3)) {
	return keyXY2;
	}else {
	return keyFit2;
	}

	case 2:
	if(state==0) {
	return keyR;
	} else if (state==1) {
	return keySelectedR;
	} else if ((state==2)\|\|(state==3)) {
	return keyXY3;
	}else {
	return keyFit3;
	}

	case 3:
	if(state==0) {
	return keyE;
	} else if (state==1) {
	return keySelectedE;
	} else if ((state==2)\|\|(state==3)) {
	return keyXY4;
	}else {
	return keyFit4;
	}

	default: return null;
	}// switch
	}// getProperty

	/**
	* update the min and max thresholds from the
	* EstimatorControl UI spinners
	* @param controlParametersMap
	*/
	public void setAllParameters(Map<String,Object> controlParametersMap) {
	if(controlParametersMap !=null) {
	maxSthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_S_KEY)).doubleValue();
	minSthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_S_KEY)).doubleValue();

	maxEthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_E_KEY)).doubleValue();
	minEthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_E_KEY)).doubleValue();

	maxIthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_I_KEY)).doubleValue();
	minIthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_I_KEY)).doubleValue();

	maxRthreshold = ((Double)controlParametersMap.get(EstimatorControl.MAX_R_KEY)).doubleValue();
	minRthreshold = ((Double)controlParametersMap.get(EstimatorControl.MIN_R_KEY)).doubleValue();

	useFreqDependantBeta = ((Boolean) controlParametersMap.get(EstimatorControl.FREQ_MODE_KEY) ).booleanValue();

	}
	return;
	}// update Thresholds



	/**
	* Estimate the parameters
	* @param localDensity
	* @param localPopulation
	* @param locationID
	* @return the ModelParameters estimate
	*/
	public abstract ModelParameters estimate(double localDensity, double localPopulation, String locationID) ;


	/**
	* Returns the raw data (state 0) and the selected data (state 1)
	* for each of the state variables S,E,I,R (chartIndex 0-3)
	* @param chartIndex
	* @param state
	* @return the data to plot
	*/
	public double[] getValues(int chartIndex, int state) {
	// pick the right data for the right chart
	if(state <= 1) {
	double[][] data = dataToPlot.get(chartIndex);
	return data[state];
	}
	// regression data
	double[][] regression = regressionValues.get(chartIndex);
	return regression[state-2];
	}// getValues


	/**
	* create a sorted regression (ordered by the x coordinate)
	* @param xList
	* @param yList
	* @return the regression
	*/
	public double[][] getSortedRegression(List<Double> xList, List<Double> yList) {
	double[][] regression = new double[2][xList.size()];
	if(yList.size() != xList.size()) return regression;
	List<PhaseSpaceCoordinate> coords = new ArrayList<PhaseSpaceCoordinate>();
	for (int i = 0; i < xList.size(); i ++) {
	double x = xList.get(i).doubleValue();
	double y = yList.get(i).doubleValue();
	PhaseSpaceCoordinate coord = new PhaseSpaceCoordinate(x,y);
	coords.add(coord);
	}
	Object[] coordArray = coords.toArray();
	Arrays.sort(coordArray);
	for (int i = 0; i < xList.size(); i ++) {
	PhaseSpaceCoordinate coord = (PhaseSpaceCoordinate)coordArray[i];
	regression[0][i] = coord.xValue;
	regression[1][i] = coord.yValue;
	}
	return regression;
	}//getSortedRegression

	}