A Machine Learning Approach to Constructing an Inflation-Themed Equity Portfolio

This post details the application of a LightGBM model to construct an inflation-themed equity portfolio. It analyzes historical equity sector and style factor behavior in different inflation environments, identifies key features (relative employment size, industry, dividend payout ratio, profit margin) that predict stock returns in high inflation periods, and describes the process of building a beta-neutral portfolio of approximately 80 stocks designed to outperform in high inflation and underperform in low inflation periods.