The new version updates the package dependencies, unit testing, and adds new distributions.

Internal update and compatibility check with scipy (distribution moment method compatible with scipy version >= 1.11.3)

Full Changelog: v1.2.0...v.1.2.1

In the Version 1.2.0 we implemented in the form of a vignette the Replication Code and the Supplementary Material of our manuscript , and updated the package with source code that did not affect the usability of the software.

What's Changed

• Drop/stretch down by @EdoLu in #7
• Severity transformed moment by @EdoLu in #8

New Contributors

• @EdoLu made their first contribution in #7

Full Changelog: v1.1.1...v1.2.0

In the new release, we made the tail extrapolation possible for the CRM.

In this new version of the package we release new functionalities for the different models.

LossModel

• Closed form moments of the aggregate loss distribution with the parameter use_dist in the methods mean, std, var, and skewness and coeff_variation.
• threshold parameter for modeling the analysis threshold of Frequency.
• In Severity, new methods: censored_mean, censored_std, censored_var, censored_skewness, and censored_coeff_variation .

LossAggregation

• New methods: lev and censored_moments.

Data sets

• New data sets in gemdata for claims reserving

LossReserve

• More explicit names for the parameters of AggregateData.
• For the CRM model it is possible to indicate the mixing frequency and mixing severity parameters as an int.

Zenodo DOI connection is created

• This version is alligned with the paper pre-print available at https://arxiv.org/abs/2303.01129.
• No new functionalities were added for this version.
• DOI created

Full Changelog: v0.3.4...v1.0.1

New plot functionalities:

• Print the discretized severity distribution c.d.f.
• Print the aggregate distribution c.d.f.

In the new release:

• The new class PolicyStructure simplified the use of LossModel .
• The new classes AggregateData and ReservingModel simplified the use of LossReserve.
• New distributions available Paralogistic, Inverse Paralogistic, Pareto1 (single-parameter Pareto) and Pareto2 (Pareto type II).
• Return type bug on LossAggregation fixed.

New features:

• lossreserve graphs adjusted and enhanced,
• gemdata development triangle naming improved,
• lossmodel aggregate deductible new parametrization,
• fundamental copulaa available, i.e. Fréchet–Hoeffding copula bounds, independent copula,
• improved external library management.