- An error related to the processing of categorical data with too many categories has been corrected.
- Removed
-march=native
flag in the build script to avoid the use of instruction sets not available on some CPUs.
- Added conditional linear Gaussian networks (
CLGNetworkType <pybnesian.models.CLGNetworkType>
,CLGNetwork <pybnesian.models.CLGNetwork>
,ConditionalCLGNetwork <pybnesian.models.ConditionalCLGNetwork>
andDynamicCLGNetwork <pybnesian.models.DynamicCLGNetwork>
). - Implemented
ChiSquare <pybnesian.learning.independences.ChiSquare>
(andDynamicChiSquare <pybnesian.learning.independences.DynamicChiSquare>
) indepencence test. - Implemented
MutualInformation <pybnesian.learning.independences.MutualInformation>
(andDynamicMutualInformation <pybnesian.learning.independences.DynamicMutualInformation>
) indepencence test. This independence test is valid for hybrid data. - Implemented
BDe <pybnesian.learning.scores.BDe>
(Bayesian Dirichlet equivalent) score (andDynamicBDe <pybnesian.learning.scores.DynamicBDe>
). - Added
UnknownFactorType <pybnesian.factors.UnknownFactorType>
as defaultFactorType <pybnesian.factors.FactorType>
for Bayesian networks when the node type could not be deduced. - Added
Assignment <pybnesian.factors.Assignment>
class to represent the assignment of values to variables.
API changes:
- Added method
Score.data() <pybnesian.learning.scores.Score.data>
. - Added
BayesianNetworkType.data_default_node_type() <pybnesian.models.BayesianNetworkType.data_default_node_type>
for non-homogeneousBayesianNetworkType <pybnesian.models.BayesianNetworkType>
. - Added constructor for
HeterogeneousBN <pybnesian.models.HeterogeneousBN>
to specify a defaultFactorType <pybnesian.factors.FactorType>
for each data type. Also, it addsHeterogeneousBNType.default_node_types() <pybnesian.models.HeterogeneousBNType.default_node_types>
andHeterogeneousBNType.single_default() <pybnesian.models.HeterogeneousBNType.single_default>
. - Added
BayesianNetworkBase.has_unknown_node_types() <pybnesian.models.BayesianNetworkBase.has_unknown_node_types>
andBayesianNetworkBase.set_unknown_node_types() <pybnesian.models.BayesianNetworkBase.set_unknown_node_types>
. - Changed signature of
BayesianNetworkType.compatible_node_type() <pybnesian.models.BayesianNetworkType.compatible_node_type>
to include the new node type as argument. - Removed
FactorType.opposite_semiparametric()
. This functionality has been replaced byBayesianNetworkType.alternative_node_type() <pybnesian.models.BayesianNetworkType.alternative_node_type>
. - Included model as argument of
Operator.opposite() <pybnesian.learning.operators.Operator.opposite>
. - Added method
OperatorSet.set_type_blacklist() <pybnesian.learning.operators.OperatorSet.set_type_blacklist>
. Added a type blacklist argument toChangeNodeTypeSet <pybnesian.learning.operators.ChangeNodeTypeSet>
constructor.
- First release! =).