Sequential asset ranking in nonstationary time series

TL;DR

This paper introduces a naive Bayes-based asset ranking algorithm that dynamically adjusts weights based on recent performance, outperforming standard benchmarks like the S&P 500.

Contribution

The paper presents a novel ranking algorithm that incorporates performance-based weight adjustments, improving asset ranking accuracy in nonstationary time series.

Findings

Outperforms long-only S&P 500 holdings

Surpasses regress-then-rank baseline

Adapts to nonstationary asset performance

Abstract

We create a ranking algorithm, the naive Bayes asset ranker. Our algorithm computes the posterior probability that individual assets will be ranked higher than other portfolio constituents. Unlike earlier algorithms, such as the weighted majority, our algorithm allows poor-performing experts to have increased weight when they start performing well. We outperform the long-only holding of the S&P 500 index and a regress-then-rank baseline.