WP-2026-001 · quant · ml
AlphaSeek - Regime-Aware Portfolio Allocation
Ibrahim Laklaa · 2026
13
market features
5
allocation methods
2009-2023
backtest window
9
pages
Abstract
Detects latent market regimes via HMM, GMM and spectral clustering, engineers 13 market features, and adapts allocation across five methods. Walk-forward backtests (2009-2023) show regime-conditional strategies beat naive baselines. 9-page LaTeX paper with auto-generated charts.
Question
Do markets move through distinct regimes, and does an allocation that knows the current regime beat one that does not?
Regimes as hidden states
Daily market features are modelled as emissions of a hidden state that follows a Markov chain:
Three detectors are compared: a hidden Markov model, a Gaussian mixture and spectral clustering, all fed with the same 13 engineered market features.
Allocation
For each detected regime, five allocation methods are evaluated. Mean-variance is the reference:
Protocol
Walk-forward backtests over 2009-2023: models are fitted on the past only, then rolled forward, to avoid look-ahead bias.
Takeaway
Regime-conditional strategies beat the naive baselines in these backtests. Full details, tables and charts are in the 9-page paper.