An Efficient Anaphora Resolution Framework for Text Summarization

View at publisher Summarise with VIRAT
Authors: Farjana Farvin Sahapudeen, D. Vaishnavi, N. Moratanch, M. Anuradha, R. Ramya
Cited by: 0
References: 26

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Anaphora resolution is a challenging discourse-level task that involves identifying the correct antecedents of referential expressions in text. This work presents a novel rule-based approach for resolving pronominal and nominal anaphora, including references to verb phrase antecedents. Unlike conventional methods that rely mainly on gender and number agreement, the proposed system incorporates high-level syntactic and semantic rules to improve resolution accuracy. These rules are integrated with fine-grained linguistic constraints reported in earlier studies, enhancing overall efficiency and robustness. Experimental results demonstrate that the proposed approach outperforms standard anaphora resolution systems by effectively utilizing rich linguistic information for antecedent identification. Anaphora resolution plays a vital role in natural language processing tasks such as term extraction, question answering, and text summarization. In this study, the proposed pronominal anaphora resolution method is applied within a cluster-based summarization framework to improve summary coherence and reduce ambiguity caused by unresolved references and achieved 4.7 ROUGE Score.

Fields of Study & Subjects

Natural Language Processing Techniques Advanced Text Analysis Techniques Topic Modeling

Publication details

Journal / source
International Conference on Sentiment Analysis and Deep Learning (ICSADL)
Pages
966–973
Publication date
18 February 2026
Language
EN