Organize causal assumptions, validation, search adjustment, portfolio construction, and reporting into one auditable sequence.
Causal-investment-management
Research Topics:
- machine-intelligence-for-finance (32)
causal-investment-management (10)
- backtesting (8)
- feature-engineering (6)
- financial-data-science (6)
- financial-modeling (3)
- time-series (3)
- factor-investing (2)
- causal-graphs (2)
- cross-validation (2)
- covariance (2)
- risk-management (2)
- dependence (2)
- probability (2)
- portfolio-optimization (2)
- optimization (2)
- research-protocol (1)
- causal-roles (1)
- confounding (1)
- identification (1)
- interference (1)
- networks (1)
- partial-identification (1)
- sensitivity-analysis (1)
- policy-evaluation (1)
- decision-rules (1)
- portfolio-choice (1)
- model-search (1)
- treatment-effects (1)
- structural-models (1)
- transportability (1)
- external-validity (1)
- deep-pdes (1)
- overfitting (1)
- validation (1)
- random-matrix-theory (1)
- decision-focused-learning (1)
- robust-optimization (1)
- deep-bsde (1)
- partial-differential-equations (1)
- deep-hedging (1)
- information-theory (1)
- sampling (1)
- financial-data (1)
- feature-importance (1)
- labeling (1)
- event-detection (1)
- fractional-differentiation (1)
- timegan (1)
- hierarchical-risk-parity (1)
- portfolio-construction (1)
- hyperparameters (1)
- model-selection (1)
- market-microstructure (1)
- liquidity (1)
- estimation (1)
- nested-clusters (1)
- financial-networks (1)
- clearing (1)
- sample-weights (1)
- overlap (1)
- majorization (1)
- portfolio-concentration (1)
- position-sizing (1)
- clustering (1)
- ensemble-learning (1)
- ambiguity (1)
- regime-change (1)
- adaptation (1)
- reproducibility (1)
- research-design (1)
- risklabai (1)
- software-design (1)
- sharpe-ratio (1)
- multiple-testing (1)
- structural-breaks (1)
- change-points (1)
- simulation (1)
- monte-carlo (1)
- tail-risk (1)
- extreme-value-theory (1)
- Read paths through a graph and distinguish common causes, mediators, colliders, and instruments before choosing an analysis.
- Show how omitted common causes and selection can make a predictive factor appear to carry a causal return effect.
- Estimate effects when one unit's treatment can change another unit's outcome through a network.
- Report the range of effects supported by the data and stated assumptions when a single point estimate is not identified.
- Estimate the value of a proposed decision rule from actions chosen by a different, recorded behavior rule.
- Carry uncertainty from factor search and model selection into the portfolio decision that follows.
- Connect potential outcomes to structural models, randomized comparisons, natural experiments, and differences over time.
- Move an estimated effect from a study population to a target population by accounting for effect modifiers and support.
- Use causal graphs to determine whether an effect can be recovered under stated assumptions.