Expert Aggregation for Financial Forecasting

TL;DR

This paper applies Bernstein Online Aggregation to combine forecasts from different machine learning models for financial time series, resulting in improved portfolio performance in non-stationary environments.

Contribution

It introduces a novel application of BOA for constructing long-short strategies from multiple models, enhancing portfolio metrics without assuming model stability.

Findings

Aggregation outperforms individual models in Sharpe Ratio

Lower shortfall achieved with aggregation

Maintains similar turnover as individual models

Abstract

Machine learning algorithms dedicated to financial time series forecasting have gained a lot of interest. But choosing between several algorithms can be challenging, as their estimation accuracy may be unstable over time. Online aggregation of experts combine the forecasts of a finite set of models in a single approach without making any assumption about the models. In this paper, a Bernstein Online Aggregation (BOA) procedure is applied to the construction of long-short strategies built from individual stock return forecasts coming from different machine learning models. The online mixture of experts leads to attractive portfolio performances even in environments characterised by non-stationarity. The aggregation outperforms individual algorithms, offering a higher portfolio Sharpe Ratio, lower shortfall, with a similar turnover. Extensions to expert and aggregation specialisations are also proposed to improve the overall mixture on a family of portfolio evaluation metrics.