OakQuant Machine Intelligence
Deterministic, statistical and generative intelligence
Clustering, peers, and changing regimes
How to group similar observations, carry uncertainty about market states, and test whether the groups improve a forecast.
Pumulo SikanetaNovember 11, 2026machine-learningclusteringmarket-regimesTemporal deep learning: when order carries the signal
How sequence models read ordered histories, what they can learn that a summary loses, and how to test them without future leakage.
Pumulo SikanetaNovember 11, 2026machine-learningdeep-learningtime-seriesPoint-in-time memory
How data vintages, availability cutoffs, and time-ordered tests keep future knowledge out of historical forecasts.
Pumulo SikanetaNovember 11, 2026machine-learningforecastingmarket-datavalidationRules engines: decisions you can replay
Explicit rules turn a request and a policy version into a repeatable decision, while exposing conflicts and missing facts.
Pumulo SikanetaNovember 11, 2026machine-learningrulesdecision-systemsBaselines and adaptive gradients
How simple forecasts set the standard, and how small supervised updates learn from each settled outcome.
Pumulo SikanetaNovember 11, 2026machine-learningforecastingadaptive-learningBayesian hierarchies and the value of a peer group
How partial pooling combines a thin company record with broader evidence without hiding uncertainty.
Pumulo SikanetaNovember 11, 2026machine-learningbayesian-modelsforecastingTrees and interactions: when one signal changes another
How decision trees learn conditional patterns, why boosted stumps can miss them, and what quantile trees add.
Pumulo SikanetaNovember 11, 2026machine-learningdecision-treesforecastingEnergy models and relationships
How energy models score joint market states, and when a relationship graph adds information beyond sector averages.
Pumulo SikanetaNovember 11, 2026machine-learningenergy-modelsforecastingCombining forecasts and calibrating confidence
How specialist forecasts become one distribution, and why event probability, return ranges, and historical skill need separate evidence.
Pumulo SikanetaNovember 11, 2026machine-learningcalibrationforecastingLearning from outcomes
How forecasts become evidence, why labels arrive late, and how a learning system changes without losing its memory.
Pumulo SikanetaNovember 11, 2026machine-learningonline-learningforecastingevaluationData provenance and cost: What did we know, and what did it cost?
How source records, receipt times, and cost-aware routing keep financial research reproducible without buying or computing more than a question needs.
Pumulo SikanetaNovember 11, 2026machine-learningmarket-datadata-qualitycost-controlGenerative capability: ideas, evidence, and words
Generative AI can propose questions and explain recorded forecasts, provided evidence and arithmetic stay outside the prose model.
Pumulo SikanetaNovember 11, 2026machine-learninglanguage-modelsforecastingexplanation