import static java.lang.System.*;

import java.util.*;

import static javastat.regression.PsiFunction.*;

import static javastat.regression.SelectionCriterion.*;

import javastat.regression.*;

import static javastat.util.Argument.*;

import javastat.util.*;

/**

*

* <p>Data Source: Brinkman, N. D., 1981. Ethanol fuel-A single-cylinder

* engine study of efficiency and exhaust emissions. SAE Transactions 90,

* No. 810345, 1410-1424. </p>

*/

public class RobustSelectionCriterionExample extends SelectionCriterionTemplate

{

public double c;

public double[] residuals;

public RobustSelectionCriterionExample(double c)

{

this.c = c;

}

public RobustSelectionCriterionExample(double c,

Hashtable argument,

Object ...dataObject)

{

this.c = c;

weightedSelectionCriterion = (Double) weightedSelectionCriterion(

argument, dataObject);

}

public RobustSelectionCriterionExample(Hashtable argument,

Object ...dataObject)

{

this(2.0, argument, dataObject);

}

public Double weightedRSS(Hashtable argument,

Object ...dataObject)

{

residuals = ((WeightedSelectionCriterion)

new WeightedSelectionCriterion(argument, dataObject).

statisticalAnalysis).residuals;

weightedRSS = 0.0;

for (int i = 0; i < residuals.length; i++)

{

if (Math.abs(residuals[i]) <= c)

{

weightedRSS += Math.pow(residuals[i], 2.0);

}

else

{

weightedRSS += 2.0 * c * Math.abs(residuals[i]) -

Math.pow(c, 2.0);

}

}

return weightedRSS;

}

public Double penalty(Hashtable argument,

Object ...dataObject)

{

return new WeightedSelectionCriterion().

penalty(argument, dataObject);

}

public static void main(String[] args)

{

DataManager dm = new DataManager();

BasicStatistics bs = new BasicStatistics();

Hashtable argument = new Hashtable();

String[][] mysqlData = new DBLoader("jdbc:mysql://localhost/data",

"root", "",

"SELECT * FROM data").data;

double[][] data = dm.transpose(dm.stringToDouble(mysqlData));

double[] response = data[0];

double[][] covariate = dm.getData(1, data.length - 1, 0,

data[0].length - 1, data);

double[][] weightMatrix =

dm.inverse(bs.covarianceAR1(response.length, 0.2, 1));

double robustAIC = (Double) new RobustSelectionCriterionExample(

argument, weightMatrix, response, covariate).

weightedSelectionCriterion;

out.println(robustAIC);

String[][] postgreData = new DBLoader(

"jdbc:postgresql://localhost/weitest", "wenwei", "1234",

"SELECT * FROM data").data;

data = dm.transpose(dm.stringToDouble(postgreData));

response = data[0];

double[] covariate2 = data[1];

argument.put(SMOOTHING_PARAMETER, 1000);

argument.put(DIVISIONS, 10);

argument.put(DEGREE, 3);

argument.put(ORDER, 2);

double robustGCV =

(Double) new RobustSelectionCriterionExample(1.5).

weightedSelectionCriterion(argument, response, covariate2);

}

}

Results:

The robust AIC is 1.329

The robust GCV is 1.126