Representation learning for regime detection in financial markets

Simple item page
dc.contributor.advisorGebbie, Timothy
dc.contributor.authorOrton, Alexa
dc.date.accessioned2025-09-18T10:53:18Z
dc.date.available2025-09-18T10:53:18Z
dc.date.issued2025
dc.date.updated2025-09-18T10:37:33Z
dc.description.abstractWe investigate financial market regime detection from the perspective of deep representation learning of the causal (reflexive) information geometry underpinning complex (multi-scale) dynamical traded asset systems using an emergent hierarchical correlation structure to characterise evolving macroeconomic market phases. Specifically, we assess the robustness of three toy models: SPD Matrix Network (SPDNet), SPD Matrix Network with Riemannian Batch Normalisation (SPDNetBN) and U-shaped SPD Matrix Network (U-SPDNet) whose architectures respect the underlying Riemannian manifold of input block hierarchical Symmetric Positive Definite (SPD) correlation matrices by employing Log-Euclidean Metric (LEM)s. Market phase detection for each model is carried out using three data configurations: i.) Randomised Johannesburg Stock Exchange (JSE) Top 60 data, ii.) synthetically-generated block hierarchical SPD matrices, and iii.) chronology-preserving block-resampled JSE Top 60 data. We show that using a singular performance metric is misleading in our financial market use cases. We confirm that U-SPDNet performs improved latent feature extraction with better classification performance in stressed and rally market phases, despite achieving lower Out-of-Sample (OOS) backtest scenario accuracy than that of the benchmark SPDNet. The SPDNet-based models fail in capturing latent reflexive spatio-temporal block hierarchical correlation dynamics and deliver corner solutions across all input data sets. The U-SPDNet is promising in terms of its utility in regime dependent portfolio optimisation strategy generation as a model better-suited to capturing latent block hierarchical correlation structures arising from lead-lag causal feedback information loops that often drive the evolution of evolving market regimes
dc.identifier.apacitationOrton, A. (2025). <i>Representation learning for regime detection in financial markets</i>. (). Universiy of Cape Town ,Faculty of Science ,Department of Statistical Sciences. Retrieved from http://hdl.handle.net/11427/41861en_ZA
dc.identifier.chicagocitationOrton, Alexa. <i>"Representation learning for regime detection in financial markets."</i> ., Universiy of Cape Town ,Faculty of Science ,Department of Statistical Sciences, 2025. http://hdl.handle.net/11427/41861en_ZA
dc.identifier.citationOrton, A. 2025. Representation learning for regime detection in financial markets. . Universiy of Cape Town ,Faculty of Science ,Department of Statistical Sciences. http://hdl.handle.net/11427/41861en_ZA
dc.identifier.ris TY - Thesis / Dissertation AU - Orton, Alexa AB - We investigate financial market regime detection from the perspective of deep representation learning of the causal (reflexive) information geometry underpinning complex (multi-scale) dynamical traded asset systems using an emergent hierarchical correlation structure to characterise evolving macroeconomic market phases. Specifically, we assess the robustness of three toy models: SPD Matrix Network (SPDNet), SPD Matrix Network with Riemannian Batch Normalisation (SPDNetBN) and U-shaped SPD Matrix Network (U-SPDNet) whose architectures respect the underlying Riemannian manifold of input block hierarchical Symmetric Positive Definite (SPD) correlation matrices by employing Log-Euclidean Metric (LEM)s. Market phase detection for each model is carried out using three data configurations: i.) Randomised Johannesburg Stock Exchange (JSE) Top 60 data, ii.) synthetically-generated block hierarchical SPD matrices, and iii.) chronology-preserving block-resampled JSE Top 60 data. We show that using a singular performance metric is misleading in our financial market use cases. We confirm that U-SPDNet performs improved latent feature extraction with better classification performance in stressed and rally market phases, despite achieving lower Out-of-Sample (OOS) backtest scenario accuracy than that of the benchmark SPDNet. The SPDNet-based models fail in capturing latent reflexive spatio-temporal block hierarchical correlation dynamics and deliver corner solutions across all input data sets. The U-SPDNet is promising in terms of its utility in regime dependent portfolio optimisation strategy generation as a model better-suited to capturing latent block hierarchical correlation structures arising from lead-lag causal feedback information loops that often drive the evolution of evolving market regimes DA - 2025 DB - OpenUCT DP - University of Cape Town KW - deep manifold representation learning KW - SPD matrix classification KW - nested block hierarchical financial market correlations KW - regime detection KW - causal feedback loops LK - https://open.uct.ac.za PB - Universiy of Cape Town PY - 2025 T1 - Representation learning for regime detection in financial markets TI - Representation learning for regime detection in financial markets UR - http://hdl.handle.net/11427/41861 ER - en_ZA
dc.identifier.urihttp://hdl.handle.net/11427/41861
dc.identifier.vancouvercitationOrton A. Representation learning for regime detection in financial markets. []. Universiy of Cape Town ,Faculty of Science ,Department of Statistical Sciences, 2025 [cited yyyy month dd]. Available from: http://hdl.handle.net/11427/41861en_ZA
dc.language.rfc3066eng
dc.publisher.departmentDepartment of Statistical Sciences
dc.publisher.facultyFaculty of Science
dc.publisher.institutionUniversiy of Cape Town
dc.subjectdeep manifold representation learning
dc.subjectSPD matrix classification
dc.subjectnested block hierarchical financial market correlations
dc.subjectregime detection
dc.subjectcausal feedback loops
dc.titleRepresentation learning for regime detection in financial markets
dc.typeThesis / Dissertation
dc.type.qualificationlevelMasters
dc.type.qualificationlevelPhD
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
thesis_sci_2025_orton alexa.pdf
Size:
6.64 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.72 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Masters