Adapting to Environmental Volatility: A Machine Learning Analysis of Stock Price Dynamics in the Korean Food Industry
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Purpose: This study investigates how external business environment dynamics influence stock price volatility in the Korean food industry and explores firms’ strategic adaptability through machine learning-based predictive modeling. Research design, data and methodology: Employing XGBoost and LSTM models, we forecast stock price movements for major Korean food companies and identify key macroeconomic and technical factors affecting volatility. Stock data were gathered via web crawling from Naver Finance, supplemented with macroeconomic indicators such as exchange rates, oil prices, and the S&P 500 index, alongside technical metrics like SMA_5 and RSI. Model performance was evaluated using R² and RMSE metrics, and feature importance analysis was conducted to uncover critical drivers. Results: XGBoost outperformed LSTM across all firms, demonstrating higher predictive accuracy and generalization. Short-term technical indicators and macroeconomic factors, particularly exchange rates and oil prices, significantly influenced stock volatility. Sector-specific variations were evident, with coffee and general food companies showing higher sensitivity to external shocks. Conclusions: The findings highlight the vital role of environmental dynamics in shaping financial performance in the food industry and underscore the utility of machine learning for strategic risk management. This study offers practical insights into how firms in the Asian food sector can better anticipate and respond to volatility induced by macroeconomic and market changes.
Abstract
Purpose: This study investigates how external business environment dynamics influence stock price volatility in the Korean food industry and explores firms’ strategic adaptability through machine learning-based predictive modeling. Research design, data and methodology: Employing XGBoost and LSTM models, we forecast stock price movements for major Korean food companies and identify key macroeconomic and technical factors affecting volatility. Stock data were gathered via web crawling from Naver Finance, supplemented with macroeconomic indicators such as exchange rates, oil prices, and the S&P 500 index, alongside technical metrics like SMA_5 and RSI. Model performance was evaluated using R² and RMSE metrics, and feature importance analysis was conducted to uncover critical drivers. Results: XGBoost outperformed LSTM across all firms, demonstrating higher predictive accuracy and generalization. Short-term technical indicators and macroeconomic factors, particularly exchange rates and oil prices, significantly influenced stock volatility. Sector-specific variations were evident, with coffee and general food companies showing higher sensitivity to external shocks. Conclusions: The findings highlight the vital role of environmental dynamics in shaping financial performance in the food industry and underscore the utility of machine learning for strategic risk management. This study offers practical insights into how firms in the Asian food sector can better anticipate and respond to volatility induced by macroeconomic and market changes.