Axiom Generator¶
Entity matching axiom generator module.
This module provides an axiom generation functionality for entity matching tasks.
- class neer_match.axiom_generator.AxiomGenerator(data_generator)¶
Axiom generator class.
The class contains the definitions for the logical connectives, functions, quantifiers, and aggregators used in the axiom generation process.
- data_generator¶
A data generator object.
- Type:
- Sim¶
A similarity predicate.
- Type:
ltn.Predicate.Lambda
- Not¶
A negation connective.
- Type:
ltn.Wrapper_Connective
- And¶
A conjunction connective.
- Type:
ltn.Wrapper_Connective
- Or¶
A disjunction connective.
- Type:
ltn.Wrapper_Connective
- Implies¶
An implication connective.
- Type:
ltn.Wrapper_Connective
- Equiv¶
An equivalence connective.
- Type:
ltn.Wrapper_Connective
- ForAll¶
A universal quantifier.
- Type:
ltn.Wrapper_Quantifier
- Exists¶
An existential quantifier.
- Type:
ltn.Wrapper_Quantifier
- FormAgg¶
A formula aggregator.
- Type:
ltn.Wrapper_Formula_Aggregator
- __init__(data_generator)¶
Initialize an axiom generator object.
By default, the axiom generator uses product t-norm and co t-norm for conjunction and disjunction, respectively. In addition, it uses Reichenbach implications. The quantifiers are based on the p-mean error and p-mean aggregators (see also https://nbviewer.org/github/logictensornetworks/logictensornetworks/blob/master/tutorials/2b-operators_and_gradients.ipynb).
- Parameters:
data_generator (
DataGenerator) – A data generator object.
- field_matching_constants()¶
Return field matching constants.
The outer list contains contains matching examples and the inner lists contain the constants of the features.
- Return type:
List[List[Constant]]
- field_non_matching_constants()¶
Return field non matching constants.
The outer list contains contains non-matching examples and the inner lists contain the constants of the features.
- Return type:
List[List[Constant]]