KIS’s new Machine Learning Factor Model uses artificial intelligence to select 10 out of 350 factors that impact stock prices of some 2,000 listed companies. (image: Kobizmedia/Korea Bizwire)
SEOUL, Jun. 28 (Korea Bizwire) — Korea Investment & Securities Co. (KIS) revealed a new South Korean stock portfolio using machine learning.
KIS’s new Machine Learning Factor Model uses artificial intelligence to select 10 out of 350 factors that impact stock prices of some 2,000 listed companies.
The model is designed to find a combination of factors that reflects the highest price earnings ratio in the last 30 days. The model is required to calculate 2.5 trillion cases to find the most optimum factor.
Based on the optimum factor models, KIS researchers have proposed 40 large capital stock items and 60 small to medium items for investment portfolios in May.
“Performance over the last month showed a 1.3 percent loss for large capital stock items, but a gain of 4.2 percent for small to medium items,” KIS said.
The model selected stockholder’s equity, gross profit margin, and cash dividend yield as major factors for large capital stock items.
For small to medium items, the model picked operating income, market capitalization, and cash flow as major factors with strongest impact on price earnings ratio.
H. M. Kang (hmkang@koreabizwire.com)
