Industry superstars: Unmasking key features that drive firm-level performance in Chinese markets using ensemble learning with genetic algorithm

Abstract

This study presents a comprehensive analysis of firm-level performance within five distinct industries, utilizing data from the Chinese Industrial Enterprises Database covering the years 2002-2007. The objective of this study is to unravel the dynamics that govern market share in given industries, with a focus on identifying key features that make a firm a “Superstar” in that industry. We designed an ensemble machine learning algorithm, with Random Forest, XGBoost, AdaBoost, and least absolute shrinkage and selection operator (LASSO) as the base learners coupled with a Genetic Algorithm (GA) for the optimal aggregation. We evaluated the sequential interplay of features influencing market share, allowing us to capture the relationships within these industries, highlighting the heterogeneity and industry-specific factors that shape market leadership. Our findings reveal that Last year’s market share consistently emerges as a significant predictor across all industries, underscoring the impact of historical performance on future market trajectory. However, the importance of other factors such as net total fixed assets and main business revenue varies across industries. This study not only contributes to the academic understanding of market dynamics but also offers practical insights for policymakers and business strategists, emphasizing the need for industry-specific approaches in decision-making.