Summarization of financial documents with TF-IDF weighting of multi-word terms

Sophie Krimberg; Natalia Vanetik; Marina Litvak

Summarization of financial documents with TF-IDF weighting of multi-word terms

Sophie Krimberg, Natalia Vanetik, Marina Litvak

Research output: Contribution to conference › Paper › peer-review

5 Scopus citations

Abstract

Financial documents, such as corporate annual reports, are usually very long and may consist of more than 100 pages. Every report is divided into thematic sections or statements that have an inner structure and include special financial terms and numbers. This paper describes an approach for summarizing financial documents based on a Bag-of-Words (BOW) document representation. The suggested solution first calculates the Term Frequency-Inverse Document Frequency (TF-IDF) weights for all single-word and multi-word expressions in the corpus, then finds the sequence of words with a maximum total weight in each document. The solution is designed to meet the requirements of the Financial Narrative Summarization (FNS 2021) shared task and has been tested on FNS 2021 dataset shared-task dataset.

Original language	English
Pages	75-80
Number of pages	6
State	Published - 1 Jan 2021
Externally published	Yes
Event	3rd Financial Narrative Processing Workshop, FNP 2021 - Lancaster, United Kingdom Duration: 15 Sep 2021 → 16 Sep 2021

Conference

Conference	3rd Financial Narrative Processing Workshop, FNP 2021
Country/Territory	United Kingdom
City	Lancaster
Period	15/09/21 → 16/09/21

ASJC Scopus subject areas

Artificial Intelligence
Business, Management and Accounting (miscellaneous)
Finance

Cite this

@conference{c10ebc2dafa146b1b0608de2b31ee1e8,

title = "Summarization of financial documents with TF-IDF weighting of multi-word terms",

abstract = "Financial documents, such as corporate annual reports, are usually very long and may consist of more than 100 pages. Every report is divided into thematic sections or statements that have an inner structure and include special financial terms and numbers. This paper describes an approach for summarizing financial documents based on a Bag-of-Words (BOW) document representation. The suggested solution first calculates the Term Frequency-Inverse Document Frequency (TF-IDF) weights for all single-word and multi-word expressions in the corpus, then finds the sequence of words with a maximum total weight in each document. The solution is designed to meet the requirements of the Financial Narrative Summarization (FNS 2021) shared task and has been tested on FNS 2021 dataset shared-task dataset.",

author = "Sophie Krimberg and Natalia Vanetik and Marina Litvak",

note = "Publisher Copyright: {\textcopyright} Proceedings of the 3rd Financial Narrative Processing Workshop, FNP 2021. All rights reserved.; 3rd Financial Narrative Processing Workshop, FNP 2021 ; Conference date: 15-09-2021 Through 16-09-2021",

year = "2021",

month = jan,

day = "1",

language = "English",

pages = "75--80",

}

TY - CONF

T1 - Summarization of financial documents with TF-IDF weighting of multi-word terms

AU - Krimberg, Sophie

AU - Vanetik, Natalia

AU - Litvak, Marina

PY - 2021/1/1

Y1 - 2021/1/1

N2 - Financial documents, such as corporate annual reports, are usually very long and may consist of more than 100 pages. Every report is divided into thematic sections or statements that have an inner structure and include special financial terms and numbers. This paper describes an approach for summarizing financial documents based on a Bag-of-Words (BOW) document representation. The suggested solution first calculates the Term Frequency-Inverse Document Frequency (TF-IDF) weights for all single-word and multi-word expressions in the corpus, then finds the sequence of words with a maximum total weight in each document. The solution is designed to meet the requirements of the Financial Narrative Summarization (FNS 2021) shared task and has been tested on FNS 2021 dataset shared-task dataset.

AB - Financial documents, such as corporate annual reports, are usually very long and may consist of more than 100 pages. Every report is divided into thematic sections or statements that have an inner structure and include special financial terms and numbers. This paper describes an approach for summarizing financial documents based on a Bag-of-Words (BOW) document representation. The suggested solution first calculates the Term Frequency-Inverse Document Frequency (TF-IDF) weights for all single-word and multi-word expressions in the corpus, then finds the sequence of words with a maximum total weight in each document. The solution is designed to meet the requirements of the Financial Narrative Summarization (FNS 2021) shared task and has been tested on FNS 2021 dataset shared-task dataset.

UR - http://www.scopus.com/inward/record.url?scp=85123300559&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85123300559

SP - 75

EP - 80

T2 - 3rd Financial Narrative Processing Workshop, FNP 2021

Y2 - 15 September 2021 through 16 September 2021

ER -

Summarization of financial documents with TF-IDF weighting of multi-word terms

Abstract

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this