models

Type Members

1. abstract class ContinuousMixtureModel[T, I, Y, YDomain, W1 <: ContinuousDistrRV[YDomain], BaseProcesses <: ContinuousProcessModel[T, I, Y, W1]] extends StochasticProcessMixtureModel[I, Y, ContinuousDistrMixture[YDomain, W1]]

   

   A process which is a multinomial mixture of continuous component processes.

   A process which is a multinomial mixture of continuous component processes.

   I

   The type of the index set (i.e. Double for time series, DenseVector for GP regression)

   Y

   The type of the output label

   W1

   Implementing class of the posterior distribution for the base processes should inherit from ContinuousMixtureRV

2. abstract class ContinuousProcessModel[T, I, Y, W <: ContinuousRandomVariable[_]] extends StochasticProcessModel[T, I, Y, W]

   

   Blueprint for a continuous valued stochastic process, abstracts away the behavior common to sub-classes such as io.github.mandar2812.dynaml.models.gp.GPRegression, io.github.mandar2812.dynaml.models.stp.StudentTRegression and others.

3. trait EvaluableModel[P, R] extends AnyRef

   

   An evaluable model is on in which there is a function taking in a csv reader object pointing to a test csv file and returns the appropriate Metrics object

   An evaluable model is on in which there is a function taking in a csv reader object pointing to a test csv file and returns the appropriate Metrics object

   P

   The type of the model's Parameters

   R

   The type of the output value

4. abstract class GenContinuousMixtureModel[T, I, Y, YDomain, YDomainVar, BaseDistr <: ContinuousDistr[YDomain] with Moments[YDomain, YDomainVar] with HasErrorBars[YDomain], W1 <: ContinuousRVWithDistr[YDomain, BaseDistr], BaseProcesses <: ContinuousProcessModel[T, I, Y, W1]] extends StochasticProcessMixtureModel[I, Y, ContMixtureRVBars[YDomain, YDomainVar, BaseDistr]]

   

   A multinomial mixture of component processes, each of which can output predictive distributions which have error bars around the mean/mode.

   A multinomial mixture of component processes, each of which can output predictive distributions which have error bars around the mean/mode.

   T

   The training data type of each component

   I

   The input feature type accepted by each component

   Y

   The type of the output label

   YDomain

   The type of a collection of outputs, e.g. vector

   YDomainVar

   The type of the second moment (variance) returned by the predictive distribution of each component process

   BaseDistr

   The type of the predictive distribution of each process.

   W1

   The random variable type returned by the predictiveDistribution() method of each component.

   BaseProcesses

   The type of the stochastic process components

5. abstract class KernelizedModel[G, L, T <: Tensor[K1, Double], Q <: Tensor[K2, Double], R, K1, K2] extends LinearModel[G, T, Q, R, L] with GloballyOptimizable with EvaluableModel[T, R]

   

6. trait LinearModel[T, P, Q, R, S] extends ParameterizedLearner[T, P, Q, R, S]

   

   Represents skeleton of a Linear Model.

   Represents skeleton of a Linear Model.

   T

   The underlying type of the data structure ex. Gremlin, Neo4j, Spark RDD etc

   P

   A Vector/Matrix of Doubles

   Q

   A Vector/Matrix representing the features of a point

   R

   The type of the output of the predictive model i.e. A Real Number or a Vector of outputs.

   S

   The type of the data containing the features and label.

7. trait Model[T, Q, R] extends AnyRef

   

   Basic Higher Level abstraction for Machine Learning models.

   Basic Higher Level abstraction for Machine Learning models.

   T

   The type of the training & test data

   Q

   The type of a single input pattern

   R

   The type of a single output pattern

8. trait ParameterizedLearner[G, T, Q, R, S] extends Model[G, Q, R]

   

   Skeleton of Parameterized Model

   Skeleton of Parameterized Model

   G

   The type of the underlying data.

   T

   The type of the parameters

   Q

   A Vector/Matrix representing the features of a point

   R

   The type of the output of the predictive model i.e. A Real Number or a Vector of outputs.

   S

   The type of the edge containing the features and label.

9. trait SecondOrderProcessModel[T, I, Y, K, M, W] extends StochasticProcessModel[T, I, Y, W]

   

   Processes which can be specified by upto second order statistics i.e.

   Processes which can be specified by upto second order statistics i.e. mean and covariance

   T

   The underlying data structure storing the training & test data.

   I

   The type of the index set (i.e. Double for time series, DenseVector for GP regression)

   Y

   The type of the output label

   K

   The type returned by the kernel function.

   M

   The data structure holding the kernel/covariance matrix

   W

   Implementing class of the posterior distribution

10. abstract class StochasticProcessMixtureModel[I, Y, W <: ContinuousMixtureRV[_, _]] extends ContinuousProcessModel[Seq[(I, Y)], I, Y, W]

    

11. trait StochasticProcessModel[T, I, Y, W] extends Model[T, I, Y]

    

    High Level description of a stochastic process based predictive model.

    High Level description of a stochastic process based predictive model.

    T

    The underlying data structure storing the training & test data.

    I

    The type of the index set (i.e. Double for time series, DenseVector for GP regression)

    Y

    The type of the output label

    W

    Implementing class of the posterior distribution

12. trait SubsampledDualLSSVM[G, L] extends KernelizedModel[G, L, DenseVector[Double], DenseVector[Double], Double, Int, Int]