Examples of com.opengamma.analytics.financial.model.volatility.smile.function.MixedBivariateLogNormalModelVolatility

Package com.opengamma.analytics.financial.model.volatility.smile.function

Examples of com.opengamma.analytics.financial.model.volatility.smile.function.MixedBivariateLogNormalModelVolatility

com.opengamma.analytics.financial.model.volatility.smile.function.MixedBivariateLogNormalModelVolatility
Data container for mixed bivariate log-normal model, where data for X,Y are reordered such that sigmasZ are appropriately ordered Derive volatility smile of a mixed log-normal model with mixed normal variable Z = X-Y under the assumption that X, Y are mixed bivariate normal variables. Here all of the parameters of X,Y and their correlations are input parameters.


    final int nData = dataStrs.length;
    final double[] res = new double[nData];
    Arrays.fill(res, 0.);


    final MixedBivariateLogNormalModelVolatility guessObjZ = new MixedBivariateLogNormalModelVolatility(wghts, sigsX, sigsY, rpfsX, rpfsY, rhos);
    final double[] sigmasZ = guessObjZ.getSigmasZ();
    final double[] relativePartialForwardsZ = guessObjZ.getRelativeForwardsZ();
    final double[] weightsZ = guessObjZ.getOrderedWeights();
    final MixedLogNormalModelData guessDataZ = new MixedLogNormalModelData(weightsZ, sigmasZ, relativePartialForwardsZ);


    for (int j = 0; j < nData; ++j) {
      final EuropeanVanillaOption option = new EuropeanVanillaOption(dataStrs[j], time, true);
      res[j] = dataVols[j] - getVolatility(option, forwardZ, guessDataZ);

View Full Code Here

    final double[][] res = new double[nData][nNormals];
    for (int i = 0; i < nData; ++i) {
      Arrays.fill(res[i], 0.);
    }


    final MixedBivariateLogNormalModelVolatility guessObjZ = new MixedBivariateLogNormalModelVolatility(wghts, sigsX, sigsY, rpfsX, rpfsY, rhos);
    final double[] sigmasZ = guessObjZ.getSigmasZ();
    final double[] relativePartialForwardsZ = guessObjZ.getRelativeForwardsZ();
    final double[] weightsZ = guessObjZ.getOrderedWeights();
    final double correction = guessObjZ.getInvExpDriftCorrection();
    final MixedLogNormalModelData guessDataZ = new MixedLogNormalModelData(weightsZ, sigmasZ, relativePartialForwardsZ);


    final double[] sigmasX = guessObjZ.getOrderedSigmasX();
    final double[] sigmasY = guessObjZ.getOrderedSigmasY();


    for (int j = 0; j < nData; ++j) {
      final EuropeanVanillaOption option = new EuropeanVanillaOption(dataStrs[j], time, true);
      final double impVolZ = getVolatility(option, forwardZ, guessDataZ);
      for (int i = 0; i < nNormals; ++i) {

View Full Code Here

    }


    rhos[0] = 0.2;
    rhos[1] = 0.9;


    final MixedBivariateLogNormalModelVolatility objTrueZ = new MixedBivariateLogNormalModelVolatility(inWeights, inSigmasX,
        inSigmasY, inRelativePartialForwardsX, inRelativePartialForwardsY, rhos);


    final double[] inSigmasZ = objTrueZ.getSigmasZ();
    final double[] inRelativePartialForwardsZ = objTrueZ.getRelativeForwardsZ();


    System.out.println("true values: " + inSigmasX[0] + "\t" + inSigmasX[1]);
    System.out.println("\n");


    for (int i = 0; i < nParams; ++i) {
      aaGuess1[i] = 1e-2 + obj.nextDouble();
    }
    MixedBivariateLogNormalFitter fitter1 = new MixedBivariateLogNormalFitter();


    boolean fitDone = false;


    while (fitDone == false) {


      for (int i = 0; i < nNorms; ++i) {
        aaGuess1X[i] = aaGuess1[i];
        aaGuess1Y[i] = aaGuess1[i + nNorms];
      }
      for (int i = 0; i < nNorms - 1; ++i) {
        aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
        aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
      }


      MixedLogNormalModelData tmpObj1X = new MixedLogNormalModelData(aaGuess1X, true);
      double[] tmpSigmasX = tmpObj1X.getVolatilities();


      System.out.println("guess: " + tmpSigmasX[0] + "\t" + tmpSigmasX[1]);
      fitter1.doFit(aaGuess1, xx, yy, time, fwdX, fwdY, nNorms, nDataPtsX, 1);
      aaGuess1 = fitter1.getParams();
      System.out.println("inintial sq: " + fitter1.getInitialSq());


      for (int i = 0; i < nNorms; ++i) {
        aaGuess1X[i] = aaGuess1[i];
        aaGuess1Y[i] = aaGuess1[i + nNorms];
      }
      for (int i = 0; i < nNorms - 1; ++i) {
        aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
        aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
      }


      tmpObj1X = new MixedLogNormalModelData(aaGuess1X, true);
      tmpSigmasX = tmpObj1X.getVolatilities();


      System.out.println("answer: " + tmpSigmasX[0] + "\t" + tmpSigmasX[1]);
      System.out.println("sq: " + fitter1.getFinalSq());
      System.out.println("\n");


      if (fitter1.getFinalSq() <= fitter1.getInitialSq() * 1e-10) {
        fitDone = true;
      } else {
        for (int i = 0; i < nParams; ++i) {
          aaGuess1[i] = 1e-2 + obj.nextDouble();
        }
        fitter1 = new MixedBivariateLogNormalFitter();
      }


    }


    for (int i = 0; i < nNorms; ++i) {
      aaGuess1X[i] = aaGuess1[i];
      aaGuess1Y[i] = aaGuess1[i + nNorms];
    }
    for (int i = 0; i < nNorms - 1; ++i) {
      aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
      aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
      aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
      aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
    }


    final MixedLogNormalModelData objAns1X = new MixedLogNormalModelData(aaGuess1X, true);
    final double[] weights = objAns1X.getWeights();
    final double[] sigmasX = objAns1X.getVolatilities();
    final double[] relativePartialForwardsX = objAns1X.getRelativeForwards();


    final MixedLogNormalModelData objAns1Y = new MixedLogNormalModelData(aaGuess1Y, true);
    final double[] sigmasY = objAns1Y.getVolatilities();
    final double[] relativePartialForwardsY = objAns1Y.getRelativeForwards();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhos);


    final double[] sigmasZ = objZ.getSigmasZ();
    final double[] relativePartialForwardsZ = objZ.getRelativeForwardsZ();


    System.out.println("Parameters XY");
    for (int i = 0; i < nNorms; ++i) {


      System.out.println(weights[i] + "\t" + inWeights[i]);
      assertEquals(weights[i], inWeights[i], Math.abs((inWeights[0] + inWeights[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasX[i] + "\t" + inSigmasX[i]);
      assertEquals(sigmasX[i], inSigmasX[i], Math.abs((inSigmasX[0] + inSigmasX[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasY[i] + "\t" + inSigmasY[i]);
      assertEquals(sigmasY[i], inSigmasY[i], Math.abs((inSigmasY[0] + inSigmasY[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsX[i] + "\t" + inRelativePartialForwardsX[i]);
      assertEquals(relativePartialForwardsX[i], inRelativePartialForwardsX[i], Math.abs((inRelativePartialForwardsX[0] + inRelativePartialForwardsX[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsY[i] + "\t" + inRelativePartialForwardsY[i]);
      assertEquals(relativePartialForwardsY[i], inRelativePartialForwardsY[i], Math.abs((inRelativePartialForwardsY[0] + inRelativePartialForwardsY[1]) / 2.) * 1e-6);
    }


    System.out.println("\n");
    System.out.println("Parameters Z");


    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasZ[i] + "\t" + inSigmasZ[i]);
      assertEquals(sigmasZ[i], inSigmasZ[i], Math.abs((inSigmasZ[0] + inSigmasZ[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsZ[i] + "\t" + inRelativePartialForwardsZ[i]);
      assertEquals(relativePartialForwardsZ[i], inRelativePartialForwardsZ[i], Math.abs((inRelativePartialForwardsZ[0] + inRelativePartialForwardsZ[1]) / 2.) * 1e-6);
    }


    System.out.println("\n");
    System.out.println("Imp Vols XYZ from fitting");


    final double[] ansVolsX = new double[100];
    final double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansVolsX[i] = volfunc.getVolatility(option, fwdX, objAns1X);
      ansVolsY[i] = volfunc.getVolatility(option, fwdY, objAns1Y);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + ansVolsX[i] + "\t" + ansVolsY[i] + "\t" + ansVolsZ[i]);
    }


    System.out.println("\n");
    System.out.println("True Imp Vols XYZ");


    final double[] trueVolsX = new double[100];
    final double[] trueVolsY = new double[100];
    final double[] trueVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      trueVolsX[i] = volfunc.getVolatility(option, fwdX, inObjX);
      trueVolsY[i] = volfunc.getVolatility(option, fwdY, inObjY);
      trueVolsZ[i] = objTrueZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + trueVolsX[i] + "\t" + trueVolsY[i] + "\t" + trueVolsZ[i]);


    }


    System.out.println("\n");
    final double[] kDataSet = {0.5, 0.7, 0.9, 1.0, 1.2, 1.5, 1.8 };
    for (final double k : kDataSet) {
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      final double trueVol = objTrueZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + trueVol);
    }


    System.out.println("\n");
    System.out.println("Imp Vols XYZ (1e-6)");


    for (int i = 0; i < 100; i++) {
      // double k = fwdX * (0.1 + 2. * i / 100.);
      assertEquals(ansVolsX[i], trueVolsX[i], Math.abs((inSigmasX[0] + inSigmasX[1]) / 2.) * 1e-6);
      assertEquals(ansVolsY[i], trueVolsY[i], Math.abs((inSigmasY[0] + inSigmasY[1]) / 2.) * 1e-6);
      assertEquals(ansVolsZ[i], trueVolsZ[i], Math.abs((inSigmasZ[0] + inSigmasZ[1]) / 2.) * 1e-6);
    }


    System.out.println("\n");
    System.out.println("Density X (1e-6)");


    final double[] ansDensityX = new double[100];
    final double[] ansDensityY = new double[100];
    final double[] ansDensityZ = new double[100];
    final double[] trueDensityX = new double[100];
    final double[] trueDensityY = new double[100];
    final double[] trueDensityZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansDensityX[i] = getDualGamma(option, fwdX, objAns1X);
      trueDensityX[i] = getDualGamma(option, fwdX, inObjX);
      //  System.out.println(k + "\t" + ansDensityX[i] + "\t" + trueDensityX[i]);
      assertEquals(ansDensityX[i], trueDensityX[i], 1e-6);
    }


    System.out.println("\n");
    System.out.println("Density Y (1e-6)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansDensityY[i] = getDualGamma(option, fwdY, objAns1Y);
      trueDensityY[i] = getDualGamma(option, fwdY, inObjY);
      //  System.out.println(k + "\t" + ansDensityY[i] + "\t" + trueDensityY[i]);
      assertEquals(ansDensityY[i], trueDensityY[i], 1e-6);
    }


    System.out.println("\n");
    System.out.println("Density Z (1e-6)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansDensityZ[i] = getDualGammaZ(option, fwdZ, objZ);
      trueDensityZ[i] = getDualGammaZ(option, fwdZ, objTrueZ);
      //  System.out.println(k + "\t" + ansDensityZ[i] + "\t" + trueDensityZ[i]);
      assertEquals(ansDensityZ[i], trueDensityZ[i], 1e-6);
    }


    System.out.println("\n");
    System.out.println("Density(exp)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.1 + 9. * i / 100.);
      final double denValueX = getDensity(weights, sigmasX, relativePartialForwardsX, k);
      final double denValueY = getDensity(weights, sigmasY, relativePartialForwardsY, k);
      final double denValueZ = getDensity(weights, sigmasZ, relativePartialForwardsZ, k);
      System.out.println(Math.log(k) + "\t" + denValueX + "\t" + denValueY + "\t" + denValueZ);
    }


    System.out.println("\n");
    System.out.println("Put-Call Parity X (1e-10)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption optionCall = new EuropeanVanillaOption(k, time, true);
      final EuropeanVanillaOption optionPut = new EuropeanVanillaOption(k, time, false);
      final double callPrice = getPrice(optionCall, fwdX, objAns1X);
      final double putPrice = getPrice(optionPut, fwdX, objAns1X);
      //     System.out.println(k + "\t" + (callPrice - putPrice) + "\t" + (fwdX - k));
      assertEquals((callPrice - putPrice), (fwdX - k), fwdX * 1e-10);
    }


    System.out.println("\n");
    System.out.println("Put-Call Parity Y (1e-10)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption optionCall = new EuropeanVanillaOption(k, time, true);
      final EuropeanVanillaOption optionPut = new EuropeanVanillaOption(k, time, false);
      final double callPrice = getPrice(optionCall, fwdY, objAns1Y);
      final double putPrice = getPrice(optionPut, fwdY, objAns1Y);
      //    System.out.println(k + "\t" + (callPrice - putPrice) + "\t" + (fwdY - k));
      assertEquals((callPrice - putPrice), (fwdY - k), fwdY * 1e-10);
    }


    System.out.println("\n");
    System.out.println("Put-Call Parity Z (1e-8)");


    for (int i = 0; i < 100; i++) {
      final double k = fwdZ * (0.01 + 2. * i / 100.);
      final EuropeanVanillaOption optionCall = new EuropeanVanillaOption(k, time, true);
      final EuropeanVanillaOption optionPut = new EuropeanVanillaOption(k, time, false);
      final double callPrice = objZ.getPriceZ(optionCall, fwdZ);
      final double putPrice = objZ.getPriceZ(optionPut, fwdZ);
      //     System.out.println(k + "\t" + (callPrice - putPrice) + "\t" + (fwdY - k));
      assertEquals((callPrice - putPrice), (fwdY - k), fwdY * 1e-10);
    }


  }

View Full Code Here


    final MixedLogNormalModelData objAns1Y = new MixedLogNormalModelData(aaGuess1Y, true);
    final double[] sigmasY = objAns1Y.getVolatilities();
    final double[] relativePartialForwardsY = objAns1Y.getRelativeForwards();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhos);


    final double[] ansVolsX = new double[100];
    final double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansVolsX[i] = volfunc.getVolatility(option, fwdX, objAns1X);
      ansVolsY[i] = volfunc.getVolatility(option, fwdY, objAns1Y);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      //   System.out.println(k + "\t" + ansVolsX[i] + "\t" + ansVolsY[i] + "\t" + ansVolsZ[i]);
    }


    final MixedBivariateLogNormalModelVolatility objTrueZ = new MixedBivariateLogNormalModelVolatility(inWeights, inSigmasX,
        inSigmasY, inRelativePartialForwardsX, inRelativePartialForwardsY, rhos);


    final double[] trueVolsX = new double[100];
    final double[] trueVolsY = new double[100];
    final double[] trueVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdY * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      trueVolsX[i] = volfunc.getVolatility(option, fwdX, inObjX);
      trueVolsY[i] = volfunc.getVolatility(option, fwdY, inObjY);
      trueVolsZ[i] = objTrueZ.getImpliedVolatilityZ(option, fwdZ);
      //     System.out.println(k + "\t" + trueVolsX[i] + "\t" + trueVolsY[i] + "\t" + trueVolsZ[i]);


    }


    final double[] inSigmasZ = objTrueZ.getSigmasZ();


    for (int i = 0; i < 100; i++) {
      // double k = fwdX * (0.1 + 2. * i / 100.);
      assertEquals(ansVolsX[i], trueVolsX[i], Math.abs((inSigmasX[0] + inSigmasX[1]) / 2.) * 1e-6);
      assertEquals(ansVolsY[i], trueVolsY[i], Math.abs((inSigmasY[0] + inSigmasY[1]) / 2.) * 1e-6);

View Full Code Here


    final MixedLogNormalModelData objAns1Y = new MixedLogNormalModelData(aaGuess1Y, true);
    final double[] sigmasY = objAns1Y.getVolatilities();
    final double[] relativePartialForwardsY = objAns1Y.getRelativeForwards();


    final MixedBivariateLogNormalModelVolatility objTrueZ = new MixedBivariateLogNormalModelVolatility(inWeights, inSigmasX,
        inSigmasY, inRelativePartialForwardsX, inRelativePartialForwardsY, rhos);


    final double[] xxZ = new double[nDataPtsX];
    final double[] yyZ = new double[nDataPtsX];
    for (int j = 0; j < nDataPtsX; ++j) {
      final EuropeanVanillaOption option = new EuropeanVanillaOption(xx[j], time, true);
      xxZ[j] = xx[j];
      yyZ[j] = objTrueZ.getImpliedVolatilityZ(option, fwdZ);
    }


    double[] rhosGuess = new double[nNorms];
    for (int i = 0; i < nNorms; ++i) {
      rhosGuess[i] = 1. - obj.nextDouble();
    }


    MixedBivariateLogNormalCorrelationFinder fitter = new MixedBivariateLogNormalCorrelationFinder();


    boolean fitRhoDone = false;
    int counterRho = 0;


    while (fitRhoDone == false) {
      ++counterRho;


      fitter.doFit(rhosGuess, xxZ, yyZ, time, weights, sigmasX, sigmasY,
          relativePartialForwardsX,
          relativePartialForwardsY, fwdX, fwdY);
      rhosGuess = fitter.getParams();
      System.out.println("\n");
      System.out.println("inintial sq: " + fitter.getInitialSq());


      System.out.println("final sq: " + fitter.getFinalSq());
      System.out.println("\n");


      if (fitter.getFinalSq() <= 1e-14) {
        fitRhoDone = true;
        System.out.println("Rho Fitting Done");
      } else {
        for (int i = 0; i < nNorms; ++i) {
          rhosGuess[i] = 1. - obj.nextDouble();
        }
        fitter = new MixedBivariateLogNormalCorrelationFinder();
      }


      ArgumentChecker.isTrue(counterRho < 500, "Too many inerations for rho. Start with new guess parameters.");
    }


    System.out.println("\n");


    rhosGuess = fitter.getParams();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhosGuess);


    final double[] ansVolsX = new double[100];
    final double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansVolsX[i] = volfunc.getVolatility(option, fwdX, objAns1X);
      ansVolsY[i] = volfunc.getVolatility(option, fwdY, objAns1Y);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      //   System.out.println(k + "\t" + ansVolsX[i] + "\t" + ansVolsY[i] + "\t" + ansVolsZ[i]);
    }


    final double[] trueVolsX = new double[100];
    final double[] trueVolsY = new double[100];

View Full Code Here

      System.out.println(yy[j] + "\t" + yyNoRand[j]);
    }


    final double[] rhos = new double[] {0.4, 0.4 };


    final MixedBivariateLogNormalModelVolatility objTrueZ = new MixedBivariateLogNormalModelVolatility(inWeights, inSigmasX,
        inSigmasY, inRelativePartialForwardsX, inRelativePartialForwardsY, rhos);


    //    final double[] inSigmasZ = objTrueZ.getSigmasZ();
    //    final double[] inRelativePartialForwardsZ = objTrueZ.getRelativeForwardsZ();


    System.out.println("true values: " + inSigmasX[0] + "\t" + inSigmasX[1]);
    System.out.println("\n");


    for (int i = 0; i < nParams; ++i) {
      aaGuess1[i] = 1e-2 + objRand.nextDouble();
    }
    MixedBivariateLogNormalFitter fitter1 = new MixedBivariateLogNormalFitter();


    boolean fitDone = false;


    while (fitDone == false) {


      for (int i = 0; i < nNorms; ++i) {
        aaGuess1X[i] = aaGuess1[i];
        aaGuess1Y[i] = aaGuess1[i + nNorms];
      }
      for (int i = 0; i < nNorms - 1; ++i) {
        aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
        aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
      }


      MixedLogNormalModelData tmpObj1X = new MixedLogNormalModelData(aaGuess1X, true);
      double[] tmpSigmasX = tmpObj1X.getVolatilities();


      System.out.println("guess: " + tmpSigmasX[0] + "\t" + tmpSigmasX[1]);
      fitter1.doFit(aaGuess1, xx, yy, time, fwdX, fwdY, nNorms, nDataPtsX, 1);
      aaGuess1 = fitter1.getParams();
      System.out.println("inintial sq: " + fitter1.getInitialSq());


      for (int i = 0; i < nNorms; ++i) {
        aaGuess1X[i] = aaGuess1[i];
        aaGuess1Y[i] = aaGuess1[i + nNorms];
      }
      for (int i = 0; i < nNorms - 1; ++i) {
        aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
        aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
        aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
      }


      tmpObj1X = new MixedLogNormalModelData(aaGuess1X, true);
      tmpSigmasX = tmpObj1X.getVolatilities();


      System.out.println("answer: " + tmpSigmasX[0] + "\t" + tmpSigmasX[1]);
      System.out.println("sq: " + fitter1.getFinalSq());
      System.out.println("\n");


      if (fitter1.getFinalSq() <= fitter1.getInitialSq() * 1e-6) {
        fitDone = true;
      } else {
        for (int i = 0; i < nParams; ++i) {
          aaGuess1[i] = 1e-2 + objRand.nextDouble();
        }
        fitter1 = new MixedBivariateLogNormalFitter();
      }


    }


    for (int i = 0; i < nNorms; ++i) {
      aaGuess1X[i] = aaGuess1[i];
      aaGuess1Y[i] = aaGuess1[i + nNorms];
    }
    for (int i = 0; i < nNorms - 1; ++i) {
      aaGuess1X[i + nNorms] = aaGuess1[i + 2 * nNorms];
      aaGuess1X[i + 2 * nNorms - 1] = aaGuess1[i + 3 * nNorms - 1];
      aaGuess1Y[i + nNorms] = aaGuess1[i + 2 * nNorms];
      aaGuess1Y[i + 2 * nNorms - 1] = aaGuess1[i + 4 * nNorms - 2];
    }


    final MixedLogNormalModelData objAns1X = new MixedLogNormalModelData(aaGuess1X, true);
    final double[] weights = objAns1X.getWeights();
    final double[] sigmasX = objAns1X.getVolatilities();
    final double[] relativePartialForwardsX = objAns1X.getRelativeForwards();


    final MixedLogNormalModelData objAns1Y = new MixedLogNormalModelData(aaGuess1Y, true);
    final double[] sigmasY = objAns1Y.getVolatilities();
    final double[] relativePartialForwardsY = objAns1Y.getRelativeForwards();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhos);


    //    final double[] sigmasZ = objZ.getSigmasZ();
    //    final double[] relativePartialForwardsZ = objZ.getRelativeForwardsZ();


    System.out.println("Parameters XY");
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(weights[i] + "\t" + inWeights[i]);
      //   assertEquals("weights ", weights[i], inWeights[i], Math.abs((inWeights[0] + inWeights[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasX[i] + "\t" + inSigmasX[i]);
      //    assertEquals("sigmasY ", sigmasX[i], inSigmasX[i], Math.abs((inSigmasX[0] + inSigmasX[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasY[i] + "\t" + inSigmasY[i]);
      //    assertEquals("sigmasY ", sigmasY[i], inSigmasY[i], Math.abs((inSigmasY[0] + inSigmasY[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsX[i] + "\t" + inRelativePartialForwardsX[i]);
      //     assertEquals("relativePartialForwardsX ", relativePartialForwardsX[i], inRelativePartialForwardsX[i], Math.abs((inRelativePartialForwardsX[0] + inRelativePartialForwardsX[1]) / 2.) * 1e-6);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsY[i] + "\t" + inRelativePartialForwardsY[i]);
      //    assertEquals("relativePartialForwardsY ", relativePartialForwardsY[i], inRelativePartialForwardsY[i], Math.abs((inRelativePartialForwardsY[0] + inRelativePartialForwardsY[1]) / 2.) * 1e-6);
    }
    //
    //    System.out.println("\n");
    //    System.out.println("Parameters Z");
    //
    //    for (int i = 0; i < nNorms; ++i) {
    //      System.out.println(sigmasZ[i] + "\t" + inSigmasZ[i]);
    //      assertEquals("sigmasZ ", sigmasZ[i], inSigmasZ[i], Math.abs((inSigmasZ[0] + inSigmasZ[1]) / 2.) * 1e-6);
    //    }
    //    for (int i = 0; i < nNorms; ++i) {
    //      System.out.println(relativePartialForwardsZ[i] + "\t" + inRelativePartialForwardsZ[i]);
    //      assertEquals("relativePartialForwardsZ ", relativePartialForwardsZ[i], inRelativePartialForwardsZ[i], Math.abs((inRelativePartialForwardsZ[0] + inRelativePartialForwardsZ[1]) / 2.) * 1e-6);
    //    }


    System.out.println("\n");
    System.out.println("Imp Vols XYZ from fitting");


    final double[] ansVolsX = new double[100];
    final double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdX * (0.1 + 2. * i / 100.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansVolsX[i] = volfunc.getVolatility(option, fwdX, objAns1X);
      ansVolsY[i] = volfunc.getVolatility(option, fwdY, objAns1Y);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + ansVolsX[i] + "\t" + ansVolsY[i] + "\t" + ansVolsZ[i]);
    }


    System.out.println("\n");
    System.out.println("True Imp Vols XYZ");

View Full Code Here


    final MixedLogNormalModelData objAns1Y = new MixedLogNormalModelData(aaGuessY, true);
    final double[] sigmasY = objAns1Y.getVolatilities();
    final double[] relativePartialForwardsY = objAns1Y.getRelativeForwards();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhos);


    final double[] sigmasZ = objZ.getSigmasZ();
    final double[] relativePartialForwardsZ = objZ.getRelativeForwardsZ();


    System.out.println("Parameters XY(weights, sigmasX, sigmasY, relfwdX, relrwdY)");
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(weights[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasX[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasY[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsX[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsY[i]);
    }


    System.out.println("\n");
    System.out.println("Parameters Z");


    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasZ[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsZ[i]);
    }


    System.out.println("\n");
    System.out.println(fwdZ);


    System.out.println("\n");
    System.out.println("Imp Vols XYZ from fitting");


    //  double[] ansVolsX = new double[100];
    //  double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdZ * (0.98 + .4 * i / 1000.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      //ansVolsX[i] = volfunc.getVolatility(option, fwdX, objAns1X);
      // ansVolsY[i] = volfunc.getVolatility(option, fwdY, objAns1Y);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      // System.out.println(k + "\t" + ansVolsX[i] + "\t" + ansVolsY[i]);
      System.out.println(k + "\t" + ansVolsZ[i]);
    }


    System.out.println("\n");

View Full Code Here


    System.out.println("\n");


    rhosGuess = fitter.getParams();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhosGuess);


    final double[] sigmasZ = objZ.getSigmasZ();
    final double[] relativePartialForwardsZ = objZ.getRelativeForwardsZ();


    System.out.println("Parameters XY(weights, sigmasX, sigmasY, relfwdX, relrwdY)");
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(weights[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasX[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasY[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsX[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsY[i]);
    }


    System.out.println("\n");
    System.out.println("fwdX: " + "\t" + fwdX);
    System.out.println("fwdY: " + "\t" + fwdY);


    System.out.println("\n");
    System.out.println("rhos");
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(rhosGuess[i]);
    }


    System.out.println("\n");
    System.out.println("Parameters Z");


    for (int i = 0; i < nNorms; ++i) {
      System.out.println(sigmasZ[i]);
    }
    for (int i = 0; i < nNorms; ++i) {
      System.out.println(relativePartialForwardsZ[i]);
    }


    System.out.println("\n");
    System.out.println("fwdZ: " + "\t" + fwdZ);


    System.out.println("\n");
    System.out.println("Imp Vols Z from fitting");


    final double[] ansVolsX = new double[100];
    final double[] ansVolsY = new double[100];
    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdZ * (0.9 + choiceOfExpiry * 0.01 + (.2 - choiceOfExpiry * 0.02) * i / 100.); //The range should be appropriately chosen depending on expiry
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, time, true);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + ansVolsZ[i]);
    }


    System.out.println("\n");
    System.out.println("Imp Vols X from fitting");

View Full Code Here


    final double[] dataStrikes = {0.8276533748061506, 0.830544004818981, 0.8356246758244018, 0.8408571903798175, 0.8438972913060586 };
    final int nData = dataStrikes.length;
    final double[] dataVols = new double[nData];


    final MixedBivariateLogNormalModelVolatility objTrueZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX, sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhosTrue);


    for (int i = 0; i < nData; ++i) {
      final EuropeanVanillaOption option = new EuropeanVanillaOption(dataStrikes[i], timeToExpiry, true);
      dataVols[i] = objTrueZ.getImpliedVolatilityZ(option, forwardZ);
    }


    final MixedBivariateLogNormalCorrelationFinder fitter = new MixedBivariateLogNormalCorrelationFinder();


    fitter.doFit(rhosGuess, dataStrikes, dataVols, timeToExpiry, weights, sigmasX, sigmasY, relativePartialForwardsX,
        relativePartialForwardsY, forwardX, forwardY);


    rhosGuess = fitter.getParams();


    final MixedBivariateLogNormalModelVolatility objFitZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhosGuess);


    final double[] fitVolsZ = new double[100];
    final double[] trueVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = forwardZ * (0.97 + .6 * i / 1000.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, timeToExpiry, true);
      fitVolsZ[i] = objFitZ.getImpliedVolatilityZ(option, forwardZ);
      trueVolsZ[i] = objTrueZ.getImpliedVolatilityZ(option, forwardZ);
      assertEquals(fitVolsZ[i], trueVolsZ[i], Math.abs((trueVolsZ[0] + trueVolsZ[99]) / 2.) * 1e-10);
    }


    for (int i = 0; i < nNormals; ++i) {

View Full Code Here


    System.out.println("\n");


    rhosGuess = fitter.getParams();


    final MixedBivariateLogNormalModelVolatility objZ = new MixedBivariateLogNormalModelVolatility(weights, sigmasX,
        sigmasY, relativePartialForwardsX, relativePartialForwardsY, rhosGuess);


    final double[] ansVolsZ = new double[100];
    for (int i = 0; i < 100; i++) {
      final double k = fwdZ * (0.97 + .6 * i / 1000.);
      final EuropeanVanillaOption option = new EuropeanVanillaOption(k, timeToExpiry, true);
      ansVolsZ[i] = objZ.getImpliedVolatilityZ(option, fwdZ);
      System.out.println(k + "\t" + ansVolsZ[i]);
    }


    System.out.println("\n");
    System.out.println(fwdZ);

View Full Code Here

TOP

Related Classes of com.opengamma.analytics.financial.model.volatility.smile.function.MixedBivariateLogNormalModelVolatility

com.opengamma.analytics.financial.model.volatility.smile.fitting.MixedBivariateLogNormalCorrelationFinder

com.opengamma.analytics.financial.model.volatility.smile.fitting.MixedBivariateLogNormalCorrelationFinderTest

com.opengamma.analytics.financial.model.volatility.smile.fitting.MixedBivariateLogNormalFitterTest

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.