Generalised Argumentation Framework

The GAF class is the abstract representation of the structure of the most relevant arguments and the relations between them for the model to be explained.

Constructor

GAF(confidence, predicted_class)

Create a GAF to be shaped by the GAFExtractor.

• confidence - confidence of the prediction

• predicted_class - class predicted by the model

Methods

add_argument(node, strength, payload)

Class method to add an argument to the graph.

add_input(node, payload)

Class method to add the input node to the graph.

add_conclusion(node, confidence, predicted_class, payload)

Class method to add the conclusion node to the graph.

remove_node(node)

Class method to remove a node from the graph.

add_relation(u, v, relation)

Class method to add a relation between two nodes (u, v).

remove_relation(u, v)

Class method to remove a relation between two nodes (u, v).

arguments()

Class method to get all arguments in the GAF.

inputs()

Class method to get all inputs in the GAF.

conclusions()

Class method to get all conclusions in the GAF.

relations_from(node)

Class method to get all relations (edges) leaving a node.

serialise(name, root_dir, payloads_dir)

Method used to serialise GAFs. This is very important for the communication between the library and the portal, as it transforms the GAFs to a JSON format and places the result in a common directory for it to be able to be processed by the portal API.

• name - name of the GAF

• root_dir - path to the directory where all visualisations are saved. Portal points to this directory as well

• payloads_dir - path to save the payloads relative to the root.