Extractive financial narrative summarisation using SentenceBERT-based clustering

Tuba Gokhan, Phillip Smith, Mark Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Downloads (Pure)

Abstract

We participate in the FNS-2021 Shared Task: “Financial Narrative Summarisation” organized by at 3rd Financial Narrative Processing Workshop (FNP-2021). We build an unsupervised extractive automatic financial summarisation system for the specific task. In our approach to the FNS-2021 shared task, the documents are first analyzed and an intermediate bespoke document set created containing the most salient sections of the reports. Next, vector representations are created for the intermediate document set based on SentenceBERT. Finally, the vectors are then clustered and sentences from each cluster are chosen for the final generated report summaries. The achieved results support the proposed method’s effectiveness.

Original languageEnglish
Title of host publicationProceedings of the 3rd Financial Narrative Processing Workshop FNP 2021
PublisherAssociation for Computational Linguistics, ACL
Pages94-98
Number of pages5
Publication statusPublished - 16 Sep 2021
Event3rd Financial Narrative Processing Workshop, FNP 2021 - Lancaster, United Kingdom
Duration: 15 Sep 202116 Sep 2021

Publication series

NameProceedings of the conference - Association for Computational Linguistics. Meeting
ISSN (Print)0736-587X

Conference

Conference3rd Financial Narrative Processing Workshop, FNP 2021
Country/TerritoryUnited Kingdom
CityLancaster
Period15/09/2116/09/21

ASJC Scopus subject areas

  • Artificial Intelligence
  • Business, Management and Accounting (miscellaneous)
  • Finance

Fingerprint

Dive into the research topics of 'Extractive financial narrative summarisation using SentenceBERT-based clustering'. Together they form a unique fingerprint.

Cite this