Package com.opengamma.strata.pricer.impl.volatility.smile

Class SmileModelFitter<T extends SmileModelData>

java.lang.Object

com.opengamma.strata.pricer.impl.volatility.smile.SmileModelFitter<T>

Type Parameters:

T - the data of smile model to be calibrated

Direct Known Subclasses:

SabrModelFitter

public abstract class SmileModelFitter<T extends SmileModelData>
extends Object

Smile model fitter.

Attempts to calibrate a smile model to the implied volatilities of European vanilla options, by minimising the sum of squares between the market and model implied volatilities.

All the options must be for the same expiry and (implicitly) on the same underlying.

Constructor Summary

Constructors

Constructor	Description
SmileModelFitter(double forward, DoubleArray strikes, double timeToExpiry, DoubleArray impliedVols, DoubleArray error, VolatilityFunctionProvider<T> model)	Constructs smile model fitter from forward, strikes, time to expiry, implied volatilities and error values.

Method Summary

Modifier and Type	Method	Description
protected Function<DoubleArray,Boolean>	getConstraintFunction(NonLinearParameterTransforms t)	Obtains the constraint function.
protected abstract DoubleArray	getMaximumStep()	Obtains the maximum number of iterations.
VolatilityFunctionProvider<T>	getModel()	Obtains the volatility function provider.
protected Function<DoubleArray,DoubleMatrix>	getModelJacobianFunction()	Obtains Jacobian function of the smile model.
protected Function<DoubleArray,DoubleArray>	getModelValueFunction()	Obtains volatility function of the smile model.
protected abstract NonLinearParameterTransforms	getTransform(DoubleArray start)	Obtains the nonlinear transformation of parameters from the initial values.
protected abstract NonLinearParameterTransforms	getTransform(DoubleArray start, BitSet fixed)	Obtains the nonlinear transformation of parameters from the initial values with some parameters fixed.
LeastSquareResultsWithTransform	solve(DoubleArray start)	Solves using the default NonLinearParameterTransforms for the concrete implementation.
LeastSquareResultsWithTransform	solve(DoubleArray start, NonLinearParameterTransforms transform)	Solve using a user supplied NonLinearParameterTransforms.
LeastSquareResultsWithTransform	solve(DoubleArray start, BitSet fixed)	Solve using the default NonLinearParameterTransforms for the concrete implementation with some parameters fixed to their initial values (indicated by fixed).
abstract T	toSmileModelData(DoubleArray modelParameters)	Obtains SmileModelData instance from the model parameters.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

SmileModelFitter

public SmileModelFitter(double forward,
                        DoubleArray strikes,
                        double timeToExpiry,
                        DoubleArray impliedVols,
                        DoubleArray error,
                        VolatilityFunctionProvider<T> model)

Constructs smile model fitter from forward, strikes, time to expiry, implied volatilities and error values.

strikes, impliedVols and error should be the same length and ordered coherently.

Parameters:

forward - the forward value of the underlying

strikes - the ordered values of strikes

timeToExpiry - the time-to-expiry

impliedVols - the market implied volatilities

error - the 'measurement' error to apply to the market volatility of a particular option TODO: Review should this be part of EuropeanOptionMarketData?

model - the volatility function provider

Method Detail

solve

public LeastSquareResultsWithTransform solve(DoubleArray start)

Solves using the default NonLinearParameterTransforms for the concrete implementation.

This returns LeastSquareResults.

Parameters:

start - the first guess at the parameter values

Returns:

the calibration results

solve

public LeastSquareResultsWithTransform solve(DoubleArray start,
                                             BitSet fixed)

Solve using the default NonLinearParameterTransforms for the concrete implementation with some parameters fixed to their initial values (indicated by fixed).

This returns LeastSquareResults.

Parameters:

start - the first guess at the parameter values

fixed - the parameters are fixed

Returns:

the calibration results

solve

public LeastSquareResultsWithTransform solve(DoubleArray start,
                                             NonLinearParameterTransforms transform)

Solve using a user supplied NonLinearParameterTransforms.

This returns LeastSquareResults.

Parameters:

start - the first guess at the parameter values

transform - transform from model parameters to fitting parameters, and vice versa

Returns:

the calibration results

getModelValueFunction

protected Function<DoubleArray,DoubleArray> getModelValueFunction()

Obtains volatility function of the smile model.

The function is defined in VolatilityFunctionProvider.

Returns:

the function

getModelJacobianFunction

protected Function<DoubleArray,DoubleMatrix> getModelJacobianFunction()

Obtains Jacobian function of the smile model.

The function is defined in VolatilityFunctionProvider.

Returns:

the function

getMaximumStep

protected abstract DoubleArray getMaximumStep()

Obtains the maximum number of iterations.

Returns:

the maximum number.

getTransform

protected abstract NonLinearParameterTransforms getTransform(DoubleArray start)

Obtains the nonlinear transformation of parameters from the initial values.

Parameters:

start - the initial values

Returns:

the nonlinear transformation

getTransform

protected abstract NonLinearParameterTransforms getTransform(DoubleArray start,
                                                             BitSet fixed)

Obtains the nonlinear transformation of parameters from the initial values with some parameters fixed.

Parameters:

start - the initial values

fixed - the parameters are fixed

Returns:

the nonlinear transformation

toSmileModelData

public abstract T toSmileModelData(DoubleArray modelParameters)

Obtains SmileModelData instance from the model parameters.

Parameters:

modelParameters - the model parameters

Returns:

the smile model data

getConstraintFunction

protected Function<DoubleArray,Boolean> getConstraintFunction(NonLinearParameterTransforms t)

Obtains the constraint function.

This is defaulted to be "unconstrained".

Parameters:

t - the nonlinear transformation

Returns:

the constraint function

getModel

public VolatilityFunctionProvider<T> getModel()

Obtains the volatility function provider.

Returns:

the volatility function provider