pflab_mean_risk | Optimal weights for one of 15 risk measures (variance, CVaR, EVaR, CDaR, max drawdown, ulcer index, Gini …) and an objective (minimum risk, maximum return, maximum utility, maximum return/risk), on any prior (sample or shrunk moments, Black-Litterman, entropy pooling, factor model, vine-copula scenarios), under bounds, budget, cardinality, group/linear constraints, turnover, tracking error, risk limits, costs and worst-case uncertainty sets. | |
pflab_efficient_frontier | The mean-risk efficient frontier for any of the 15 risk measures and any prior and constraints: points Pareto-optimal portfolios from minimum risk to maximum return, or one per target return or target risk level (unreachable targets are dropped with a warning). | |
pflab_risk_budgeting | Risk budgeting: weights whose contributions to a risk measure (variance, CVaR, CDaR, max drawdown … any of the 15) are proportional to risk_budget (equal by default: risk parity). | |
pflab_max_diversification | The maximum diversification portfolio: weights maximising the diversification ratio (weighted average of asset volatilities over portfolio volatility), under any constraints and prior. | |
pflab_naive | Naive allocations used as baselines: equal weight (1/N), inverse volatility (from any prior’s covariance) or a random Dirichlet draw. | |
pflab_stacking | Stacking: several base strategies (any optimizer, prior and constraints) are fitted, their out-of-sample returns (by cross-validation) become the assets of a final optimizer, and the final asset weights are the stack-weighted mix of the base weights. | |
pflab_robust_cvar | Distributionally robust mean-CVaR: maximises expected return minus risk_aversion x CVaR under the worst distribution within a Wasserstein ball of the scenarios, which guards against estimation error in the return distribution. | |
pflab_benchmark_tracker | The portfolio of the given assets that best tracks a benchmark: minimises the risk (tracking error by default) of the portfolio’s return minus the benchmark’s, under any constraints. | |
