Aydınhan Kolm Mulvey Shu 2024
“Identifying patterns in financial markets: extending the statistical jump model for regime identification” was published in Annals of Operations Research (Springer) by Afşar Onat Aydınhan, Petter N. Kolm, John M. Mulvey and Yizhan Shu (SSRN abstract 4556048, dated 20 March 2024). It extends the Statistical Jump Model into a continuous-state variant — the continuous statistical jump model (CJM). The discrete jump model, introduced by Nystrup Lindström Madsen 2020 and developed in Nystrup Kolm Lindström 2021, assigns each period a single hard regime label by clustering temporal features while penalising jumps across regimes. The CJM instead generalises the model so each period carries a probability vector over all possible regimes, allowing smoother and more dynamic transitions than the rigid classification of the discrete model.
The paper’s two methodological contributions are the continuous-state representation itself and a novel “mode loss” penalty, which the authors report refines regime identification specifically under conditions of regime imbalance and limited data — a recurring practical problem because bear and crisis regimes are short and rare relative to bull regimes. The CJM is one of the variants shipped in the open-source jumpmodels Python library, so the method is reproducible; this is why replication_available is set to yes even though no independent replication of a trading result exists.
It appears in this vault as the second-generation methodology extension of the jump-model family and as evidence of active development of the model. Crucially, it is not an independent costed out-of-sample trading study: it is framed for “regime-aware portfolio management and risk assessment,” and its empirical numbers are diagnostic regime-identification results rather than a transaction-cost-net backtest with a benchmark, Sharpe and drawdown. Because its authorship overlaps with both the originating Nystrup/Kolm jump-model papers (Kolm) and the Princeton downside-risk trading study (Mulvey, Shu — see Shu Yu and Mulvey 2024), it does not constitute independent cross-group confirmation of a tradeable edge. Its profitability grade is therefore inconclusive: the CJM is a genuine, reproducible methodological advance for Regime Classification, but the vault has no evidence that the continuous representation produces net-of-cost alpha.
Connections
- Statistical Jump Model — proposes_model, 2024, source: https://doi.org/10.1007/s10479-024-06035-z
- Hidden Markov Model Regime Detection — compares_benchmark, source: https://doi.org/10.1007/s10479-024-06035-z
- Regime Classification — detects_regime, source: https://doi.org/10.1007/s10479-024-06035-z
- Nystrup Lindström Madsen 2020 — relates, source: https://doi.org/10.1007/s10479-024-06035-z
- Nystrup Kolm Lindström 2021 — relates, source: https://doi.org/10.1007/s10479-024-06035-z
- Petter Kolm — proposes_model, source: https://doi.org/10.1007/s10479-024-06035-z
- John Mulvey — proposes_model, source: https://doi.org/10.1007/s10479-024-06035-z
- Shu Yu and Mulvey 2024 — relates, source: https://doi.org/10.1007/s10479-024-06035-z
Ontology
Aydınhan Kolm Mulvey Shu 2024 defines Statistical Jump Model Aydınhan Kolm Mulvey Shu 2024 supports Regime Classification Aydınhan Kolm Mulvey Shu 2024 relates Hidden Markov Model Regime Detection
Sources
- Aydınhan, A. O., Kolm, P. N., Mulvey, J. M. & Shu, Y. (2024). “Identifying patterns in financial markets: extending the statistical jump model for regime identification.” Annals of Operations Research. https://doi.org/10.1007/s10479-024-06035-z
- SSRN preprint (abstract id 4556048, 20 March 2024). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4556048