Bank of Korea Holds Joint Forum with Statistical Society: "Economic Statistics Evolving through AI Utilization"

Yoo Sangdae: "AI Utilization Expanding in Statistical Research"

On September 19, the Bank of Korea held a joint forum with the Korean Statistical Society under the theme "The Evolution of Economic Statistics: Expanding Statistical Methodologies and the Use of AI."

The event brought together practitioners and analytical experts in statistics from academia, related institutions, and the Bank of Korea to present research findings and share knowledge. At this forum, the two organizations shared research results on expanding methodologies for data analysis and statistical compilation using artificial intelligence (AI) and big data, and explored future challenges.

Yoo Sangdae, Deputy Governor of the Bank of Korea, attended the forum and stated, "Recently, the pace of AI development has accelerated, and the use of artificial intelligence is spreading across various fields. The Bank of Korea and the Korean Statistical Society chose the topic of this forum to keep pace with the times."

He added, "Over the past decade, we have shared a variety of topics through our joint forums, consistently discussing the digital economy, artificial intelligence, and big data. I hope that the research findings and ideas presented today will be widely used in academia and policy fields, greatly contributing to our understanding of and response to the economy."

The forum, which continues until the afternoon, will feature three sessions: invited lectures, research presentations, and discussions.

In Session 1, Oh Heeseok, Professor of Statistics at Seoul National University, will give a lecture on the "data-adaptive factor model," which can improve the accuracy of economic and financial data analysis and forecasting. Professor Oh will present a statistical model that can more precisely capture idiosyncratic volatility structures and enhance prediction accuracy for indicators such as stock returns and unemployment rates.

Session 2 will explore new methodologies for explaining and forecasting economic indicators using diverse information sources, such as news data, images, and stock market supply-demand trends, in addition to traditional economic indicators. Seo Beomseok, Professor of Statistics at Sookmyung Women's University, will propose a methodology to improve exchange rate forecasting by utilizing multiview data, including not only macroeconomic data but also exchange rate trend images and news data. Han Heejun, Professor of Statistics at Sungkyunkwan University, will introduce a new neural network-based model (Multi-Input LSTM) to improve the accuracy of stock index volatility forecasts.

Session 3 will present statistical research achievements from the Bank of Korea. Kim Sojeong, Manager of the Bank of Korea's Statistical Research Team, will introduce a new methodology for calculating the News Sentiment Index using a Small Language Model (SLM) to enhance its utility and reliability. Unlike the existing index, which was based on certain sentences, the new approach uses SLM to learn from the entire body of economic news articles. In addition, Kim Kiyong, Manager of the Statistical Information Team, and an investigator from the Input-Output Team, will share results from estimating Korea's data production value using machine learning techniques. When including the output of data and databases, Korea's 2020 Gross Domestic Product (GDP) increased by 56.3 trillion won, accounting for 3% of GDP.

This content was produced with the assistance of AI translation services.