Module large_losses 

Large-loss data with the two usual defects, and maximum likelihood fits of Pareto-family severities to it (see docs/design/pareto.md).

• Reporting thresholds: loss i was only recorded because it exceeded r_i, so its likelihood is conditional on X > r_i.
• Censoring: a loss capped by a policy limit is known only to be at least its recorded value, and contributes P(X ≥ y_i | X > r_i).

Weights count a loss several times (or a fraction of a time).

Structs

LargeLosses

Large losses for maximum likelihood fits: values, and optionally a reporting threshold, a censoring flag and a weight per loss.