Measuring News Sentiment of Korea Using Transformer

Beomseok Seo, Younghwan Lee and Hyungbae Cho
Additional contact information
Beomseok Seo: Bank of Korea
Younghwan Lee: Bank of Korea
Hyungbae Cho: Bank of Korea

Korean Economic Review, 2024, vol. 40, 149-176

Abstract: We have developed the Korean news sentiment index (NSI) to gauge the economic sentiment of Korea on a daily basis by analyzing news texts gathered from the Internet. Our framework utilizes cutting-edge natural language processing techniques to compute the NSI and examine keywords, offering insights into its fluctuations. We designed a sentiment classifier using transformer neural networks that effectively process extensive news samples to compute the NSI of Korea. We compute the NSI more frequently and immediately than official indices that rely on monthly surveys. Through this, we can identify changes in economic sentiment before official statistics are released. Moreover, the proposed framework offers keyword analysis and sector indices to clarify why economic sentiments fluctuate. Our comprehensive assessments demonstrate that the NSI is a valuable leading index and an essential tool for identifying inflection points in economic sentiment.

Keywords: News Text Data; Natural Language Processing for Economic Analysis; Sentiment Shocks (search for similar items in EconPapers)
JEL-codes: C45 C82 E32 (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://keapaper.kea.ne.kr/RePEc/kea/keappr/KER-20240101-40-1-05.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:kea:keappr:ker-20240101-40-1-05

Access Statistics for this article

Korean Economic Review is currently edited by Kyung Hwan Baik

More articles in Korean Economic Review from Korean Economic Association Contact information at EDIRC.
Bibliographic data for series maintained by KEA ().