Rules

The galactic.rules module.

It implements basic classes for dealing with bases of rules:

Rule basic class for representing implications.
Basis represents all rule bases;
BasicBasis represents bases storing their rules;
GeneratedBasis represents bases generating their rules;
CanonicalDirectBasis represents canonical direct bases;
HasseDiagramOrderedBasis represents Hasse diagram ordered bases;
SimplifiedBasis represents simplified bases.
RuleTable for displaying rule bases in jupyter notebook.

class Rule(premises: Iterable[Predicate], conclusions: Iterable[Predicate], support: SupportsFloat = 1.0, confidence: SupportsFloat = 1.0)

The Rule class.

It is used to implement an implication between premises and conclusions.

The support of a rule $A \rightarrow B$ is equal to the number of objects that match the rule divided by the total number of objects.

The confidence of a rule $A \rightarrow B$ is equal to the support of $A \cup B$ divided by the support of $A$.

__init__(premises: Iterable[Predicate], conclusions: Iterable[Predicate], support: SupportsFloat = 1.0, confidence: SupportsFloat = 1.0)

Initialise a Rule instance.

Parameters:

premises (Iterable[Predicate]) – An iterable of premises.
conclusions (Iterable[Predicate]) – An iterable of conclusions.

Keyword Arguments:

support (SupportsFloat) – The rule support.
confidence (SupportsFloat) – The rule confidence.

classmethod __init_subclass__(**kwargs: Any): Add a sub-class hook.

conclusions() → Iterator[Predicate]

Get an iterator over the conclusions.

Returns:: An iterator over the conclusions.
Return type:: Iterator[Predicate]

property confidence: float

Get the confidence.

Returns:: The confidence.
Return type:: float

premises() → Iterator[Predicate]

Get an iterator over the premises.

Returns:: An iterator over the premises.
Return type:: Iterator[Predicate]

property support: float

Get the support.

Returns:: The support.
Return type:: float

class Basis

The Basis class is the root of all implicational basis.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

class BasicBasis(rules: Iterable[Rule])

The BasicBasis class is used to construct implicational basis.

__init__(rules: Iterable[Rule])

Initialise a BasicBasis instance.

Parameters:: rules (Iterable[Rule]) – Iterable of rules.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

class GeneratedBasis(lattice: ConceptLattice, prefix: str = '$')

The GeneratedBasis is the super class of all generated basis.

__init__(lattice: ConceptLattice, prefix: str = '$')

Initialise a GeneratedBasis instance.

Parameters:: lattice (ConceptLattice) – The concept lattice.
Keyword Arguments:: prefix (str) – The prefix used for identifiers.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

property lattice: ConceptLattice

Get the concept lattice.

Returns:: The concept lattice.
Return type:: ConceptLattice

property prefix: str

Get the prefix.

Returns:: The prefix.
Return type:: str

class CanonicalDirectBasis(lattice: ConceptLattice, prefix: str = '$')

The CanonicalDirectBasis is used to manage the canonical direct basis.

__init__(lattice: ConceptLattice, prefix: str = '$')

Initialise a GeneratedBasis instance.

Parameters:: lattice (ConceptLattice) – The concept lattice.
Keyword Arguments:: prefix (str) – The prefix used for identifiers.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

property lattice: ConceptLattice

Get the concept lattice.

Returns:: The concept lattice.
Return type:: ConceptLattice

property prefix: str

Get the prefix.

Returns:: The prefix.
Return type:: str

class HasseDiagramOrderedBasis(lattice: ConceptLattice, prefix: str = '$')

The HasseDiagramOrderedBasis class.

It is used to construct implicational basis from a concept lattice using the links from the Hasse diagram.

__init__(lattice: ConceptLattice, prefix: str = '$')

Initialise a GeneratedBasis instance.

Parameters:: lattice (ConceptLattice) – The concept lattice.
Keyword Arguments:: prefix (str) – The prefix used for identifiers.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

property lattice: ConceptLattice

Get the concept lattice.

Returns:: The concept lattice.
Return type:: ConceptLattice

property prefix: str

Get the prefix.

Returns:: The prefix.
Return type:: str

class SimplifiedBasis(rules: Iterable[Rule], reduction: bool = False, generalization: bool = False, composition: bool = False)

The SimplifiedBasis class is used to manage simplified basis of rules.

It uses axioms from Malaga’s team.

__init__(rules: Iterable[Rule], reduction: bool = False, generalization: bool = False, composition: bool = False)

Initialise a Simplified instance.

Parameters:

rules (Iterable[Rule]) – An iterable of rules.
reduction (bool) – The reduction flag.
generalization (bool) – The generalization flag.
composition (bool) – The composition flag.

filter(support: Optional[SupportsFloat] = None, confidence: Optional[SupportsFloat] = None) → Iterator[Rule]

Get the rules with a minimum support and minimum confidence.

Keyword Arguments:

support (SupportsFloat, optional) – The minimum support
confidence (SupportsFloat, optional) – The minimum confidence

Returns:

An iterator over the rules.

Return type:

Iterator[Rule]

class RuleTable(basis: Basis, premise_width: int = 1, conclusion_width: int = 1, support_width: int = 1, confidence_width: int = 1, support: Optional[float] = None, confidence: Optional[float] = None)

The Table class is used to represent lattice in jupyter notebooks.

__init__(basis: Basis, premise_width: int = 1, conclusion_width: int = 1, support_width: int = 1, confidence_width: int = 1, support: Optional[float] = None, confidence: Optional[float] = None) → None

Initialise a Table class.

Parameters:

basis (Basis) – The implicational basis whose table representation is requested

Keyword Arguments:

premise_width (int) – The premise width.
conclusion_width (int) – The conclusion width.
support_width (int) – The support width.
confidence_width (int) – The confidence width.
support (Optional[float]) – The minimal support
confidence (Optional[float]) – The minimal confidence.

The values of the given widths are compared with each other to allocate relative column sizes.

property basis: Basis

Get the underlying basis.

Returns:: The underlying basis.
Return type:: Basis

property conclusion_width: int

Get the conclusion width.

Returns:: The conclusion width.
Return type:: int

property confidence: Optional[float]

Get the confidence.

Returns:: The confidence.
Return type:: Optional[float]

property confidence_width: int

Get the confidence width.

Returns:: The confidence width.
Return type:: int

property premise_width: int

Get the premise width.

Returns:: The premise width.
Return type:: int

property support: Optional[float]

Get the support.

Returns:: The support.
Return type:: Optional[float]

property support_width: int

Get the support width.

Returns:: The support width.
Return type:: int

register_classes() → None: Register all classes defined in this module.