jaclearn.logic.decision_tree.rule#

Classes

`AtomicDecisionRule`
`DecisionRule`	A decision rule is a pair of (DNF, label).
`DecisionRuleFormat`

Functions

extract_rule(decision_tree, feature_names[, ...])

Extract logic rules (DNF) from a trained DecisionTreeClassifier.

Class AtomicDecisionRule

class AtomicDecisionRule[source]#

Bases: object

__init__(variable, threshold, right_branch=False)[source]#

Instantiate an atomic decision rule. That is, by comparing a variable with a threshold.

left branch: variable <= threshold
right branch: variable > threshold

Parameters:

variable (str) – the name for the variable.
threshold (Optional[float]) – the threshold. If set to none, indicates that the variable is a boolean variable.
right_branch (Optional[bool]) – the boolean indicator.

__new__(**kwargs)#

format(format)[source]#

Format an atomic decision rule into a string.

Parameters:: format (Union[DecisionRuleFormat, str]) – the format.

Class DecisionRule

class DecisionRule[source]#

Bases: object

A decision rule is a pair of (DNF, label).

__init__(clauses, label, probabilities=None)[source]#

Instantiate a decision formula.

Parameters:

clauses (Sequence[Sequence[AtomicDecisionRule]]) – the DNF, represented as a two-level nested list.
label (Any) – the output label.
probabilities (List[float] | None)

__new__(**kwargs)#

format_clause(format)[source]#

Parameters:: format (DecisionRuleFormat | str)

Class DecisionRuleFormat

class DecisionRuleFormat[source]#

Bases: JacEnum

__new__(value)#

classmethod assert_valid(value)#: Assert if the value is a valid choice.

classmethod choice_names()#: Returns the list of the name of all possible choices.

classmethod choice_objs()#: Returns the list of the object of all possible choices.

classmethod choice_values()#: Returns the list of the value of all possible choices.

classmethod from_string(value)#

Parameters:: value (str | JacEnum)
Return type:: JacEnum

classmethod is_valid(value)#: Check if the value is a valid choice.

classmethod type_name()#: Return the type name of the enum.

LISP = 'lisp'#

PYTHON = 'python'#

Functions

extract_rule(decision_tree, feature_names, boolean_input=True, boolean_output=False, multi_output=False)[source]#

Extract logic rules (DNF) from a trained DecisionTreeClassifier.

Parameters:

decision_tree (DecisionTreeClassifier) – a pre-trained decision tree.
feature_names (Sequence[str]) – a list of strings for the features.
boolean_input (Optional[bool]) – whether the input features are boolean or not.
boolean_output (Optional[bool]) – whether the output features are boolean or not.
multi_output (bool) – whether the output features are multi-valued or not. In this case, all possible values at a leaf node will be registered.

Returns:

A mapping from label (strings, integers…) to the corresponding rule.

Return type:

rule_dict (Mapping[Any, DecisionRule])