The work presented describes a combined approach to the partial extraction of the argumentative structure of a text that can be employed in the absence of sufficient annotated data to apply efficiently the machine learning methods for the direct detection of arguments and their relations. In this case, argument identification is performed by using the patterns of argumentation indicators created by a linguist and automatically expanded. These patterns enable the recognition of specific argument types with fine precision. In this study, arguments “from expert opinion” serve as such a pivot type. Besides, potential relations between recognized arguments are analyzed by dividing the text into superphrasal units (fragments united by one topic). The criterion for connecting arguments in an argumentative structure is their inclusion in the same superphrasal unit. An experiment for identifying potentially related arguments is conducted on a set of popular science texts with a minimum size of 1000 words.
The presented work describes the analysis of argumentative statements included into the same text topic fragment as a recognition feature in terms of its efficiency. This study is performed with the purpose of using this feature in automatic recognition of argumentative structures presented in the popular science texts written in Russian. The topic model of a text is constructed based on superphrasal units (text fragments united by one topic) that are identified by detecting clusters of words and word-combinations with the use of scan statistics. Potential relations, extracted from topic models, are verified through the use of texts with manually annotated argumentation structures. The comparison between potential (based on topic models) and manually constructed relations is performed automatically. Macro-average scores of precision and recall are equal to 48.6% and 76.2% correspondingly.
В статье представлены результаты исследования изменений, происходящих в тематических кластерах, построенных на коллекции текстов конференций предметной области Argument mining. Выявление терминов, установление связей между ними и тематическая кластеризация проведены с помощью сторонних программных средств, позволяющей извлекать термины в форме именных словосочетаний, проводить их кластеризацию на базе алгоритма, основанного на применении функции модулярности. Приводится оценка качества полученных кластеров по трем критериям. Трансформацию терминологического состава кластеров во времени предлагается анализировать с помощью ориентированных графов, построенных на основе критерия, который позволяет фиксировать наиболее важные изменения.Терминологическая лексика выявленных тематических кластеров характеризует отдельные направления, в которых ведутся исследования, а трансформация терминологического состава кластеров во времени демонстрирует смещение интересов.
The study aims to determine the compatibility of arguments from different functional groups in a collection of scientific texts. The study is novel in that it develops a functional classification of argumentation schemes and identifies the features of using arguments from different functional groups in the collection of Russian-language scientific texts (it is the first time that such an analysis of functional compatibility of arguments has been carried out both for texts of the scientific genre and for texts in Russian). Based on a comparative analysis of the semantics of arguments and the functional features of their use, a classifi-cation of argumentation schemes has been developed differentiating four methods of proof (from authority, from practical value, through elaboration or causal analysis). The use of arguments from four groups has been investigated using a set of 1030 reasoning sequences extracted from expertly annotated scientific papers on linguistics and computer technology. It has been shown that the analysed papers are characterised by an active combination of arguments from different functional groups with their uneven positional arrangement in some sequences, depending on the emphasis in the proof. The work includes the following parts: argumentation modelling, a functional comparison of argumentation schemes, presentation of reasoning through functional blocks, a compatibility analysis of such arguments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.