Empirical Asset Pricing via Gaussian Process Regression
Date: 8 June 2022 / 09:00 - 17:00
Room D1.14, 16:30-17:30
We introduce an ensemble learning method based on Gaussian Process Regression (GPR) for predicting expected stock returns conditional on stock-level and macro-economic features. Our ensemble learning approach significantly reduces the computational complexity inherent in GPR inference. We perform an empirical analysis on a large cross-section of US stocks from 1962 to 2016. We find that our method dominates existing machine learning models both statistically and economically, in terms of out-of-sample R2 and Sharpe ratios of prediction-sorted portfolios, respectively. Exploiting the Bayesian nature of GPR, we also construct the optimal mean-variance and minimum variance portfolios based on the posterior covariance matrix. Both perform significantly better than the S&P 500 and equally-weighted portfolio.