A Framework for Detection and Identification the Components of Arguments in Arabic Legal Texts

Abstract

Argument mining processes in the legal domain it is aiming to detect and extract the premises, claims and their relations automatically from unstructured legal texts to provide structured data that can be processable by argumentation models. This paper presents a framework to detect and identify the components of arguments in texts of Arabic legal documents. The framework proposes a computational model adopts the supervised learning that integrates an annotated Arabic Legal Text corpus (ALTC), it is a collection of Iraq's Federal Court of Cassation decision documents with different binary classifiers based on relevant features to detect and identify the components of arguments from legal decisions texts as a final goal. The results of the framework experiments are promising, especially this paper is the first in argumentation mining processing at the level of the Arabic texts.