mlpy.mdp.MDPModelFactory.create

static MDPModelFactory.create(_type, *args, **kwargs)[source]

Create an MDP model of the given type.

Parameters:

_type : str

The MDP model type. Valid model types:

discretemodel

A model for discrete state and actions deriving transition and reward information from empirical data. A DiscreteModel instance is created.

decisiontreemodel

A model for discrete state and actions deriving transition and reward information from empirical data generalized using decision trees. A DecisionTreeModel instance model is created.

casml

A model for continuous state and actions deriving transition information from empirical data fit to a case base and a Hidden Markov Model (HMM). Rewards are derived from empirical data. A CASML instance is created.

args : tuple, optional

Positional arguments to pass to the class of the given type for initialization.

kwargs : dict, optional

Non-positional arguments to pass to the class of the given type for initialization.

Returns:

IMDPModel :

A MDP model instance of the given type.