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 using data from the Chinese Industrial Enterprises Database. We designed an ensemble machine learning algorithm (Random Forest, XGBoost, AdaBoost, LASSO) coupled with a Genetic Algorithm for optimal aggregation to identify factors driving market share and ‘Superstar’ status in various industries. We highlight the consistency of historical performance as a predictor along with heterogeneity in the importance of fixed assets and revenue across industries.

Mohammad Fili

Postdoctoral Research Fellow

My research interests include Healthcare Data Analytics, Machine Learning, and Optimization.