Actuarial reserving in Python
MPL-2.0 License
Bot releases are visible (Hide)
Published by jbogaardt over 3 years ago
groupby
parameter that allows for apriroi computation at user-specified grain for granular trianglesgroupby
support in key development estimatorscum_to_incr()
not working as expected on full_triangle_
full_triangle_
cdf broadcasting bugsigma_
or std_err_
properties after tail fitPublished by jbogaardt over 3 years ago
.at
and .iat
slicing and value assignmentfull_triangle_
run-off for expected loss methods.Published by jbogaardt over 3 years ago
valuation_date
bug occuring when partial year Triangle is instantiated as a vector.sparse>=0.12.0
Published by jbogaardt over 3 years ago
truncation_age
in the TailClark
estimator to replicate examples from the paperTweedieGLM
and DevelopmentML
to take advantage of a broader set of (sklearn-compliant) regression frameworks.PatsyFormula
to make working with ML algorithms easier.IncrementalAdditive
method comply with the rest of the package APIPublished by jbogaardt over 3 years ago
groupby
support.reg_threshold
argument to TailCurve
for finer control of fit. Huge thanks to @brian-13 for the contribution.VotingChainladder
workflow estimator to allow for averaging multiple IBNR estimators Huge thanks to @cbalona for the contribution.CaseOutstanding
development estimatorCapCod
functionality to that of Benktander
and BornhuetterFerguson
grain
issues when using the trailing
argumentv0.7.12
Trend
estimator APIBerquistSherman
estimatorPipeline
xlcompose
. Will be removed as a depdendency in v0.9.0Published by jbogaardt over 3 years ago
No code changes from 0.7.11
. Bump release to fix conda packaging.
Published by jbogaardt over 3 years ago
virtual_column
functionalitysort_index
not accepting kwargsDevelopmentConstant
when using a callablePublished by jbogaardt over 3 years ago
virtual_column
functionalitysort_index
not accepting kwargsDevelopmentConstant
when using a callablePublished by jbogaardt almost 4 years ago
sample_weight
error handling on predict - thank you @cbalonaloc
consistency with pandaspd.Series
Published by jbogaardt almost 4 years ago
ParallelogramOLF
transformerTrend
transformerPublished by jbogaardt almost 4 years ago
to_frame
now works on Triangles that are 3D or 4D. For example clrd.to_frame()
groupby
operations supported. For (trivial) example:clrd = cl.load_sample('clrd')
# Split companies with names less than 15 characters vs those above:
clrd.groupby(clrd.index['GRNAME'].str.len()<15).sum()
Published by jbogaardt about 4 years ago
Development
can now take a float value for averaging. When float value is used, it corresponds to weight exponent (delta in Barnett/Zenwirth). Only special cases had previously existed - {"regression": 0.0, "volume": 1.0, "simple": 2.0}
n_periods
with asymmetric triangles fixedPublished by jbogaardt about 4 years ago
clrd = cl.load_sample('clrd')
clrd.iloc[[0,10, 3], 1:8, :5, :]
clrd.loc[:'Aegis Grp', 'CumPaidLoss':, '1990':'1994', :48]
origin_as_datetime
argument for better compatibility with various plotting libraries (Thank you @johalnes )tri.to_frame(origin_as_datetime=True)
sparse
and numpy
arrays to accomodate more scenarios.latest_diagonal
for sparse backendsMackChainladder
which errored on asymmetric triangles (Thank you @johalnes for reporting)Published by jbogaardt about 4 years ago
clrd = cl.load_sample('clrd')
clrd / clrd.groupby('LOB').sum() # LOB alignment works now instead of throwing error
Published by jbogaardt about 4 years ago
Tiny release.
head
and tail
methods to TrianglePublished by jbogaardt over 4 years ago
valuation
axisgroupby
sort_index
method to Triangle
consistent with pandasfit_predict
to be called on a Pipeline
estimatorPublished by jbogaardt over 4 years ago
predict
and fit_predict
methods added to all IBNR models and now function as expectedDevelopmentConstant
to take on more than one set of patterns by passing in a callableMunichAdjustment
Allow ` does not work when P/I or I/P ratios cannot be calculated. You can now optionally back-fill zero values with expectaton from simple chainladder so that Munich can be performed on sparser triangles.val_to_dev
ldf_
, sigma
, and std_err_
trianglescdf_
, full_triangle_
, full_expectation_
, ibnr_
to function-based properties instead of in-memory objects to reduce memory footprintPublished by jbogaardt over 4 years ago
link_ratio
outliersTriangle.trend
to_json
. Required addition for the new BerquistSherman
methodMunichAdjustment.transform
BootstrapODPSample
to workflow section as it is not a development estimator.Published by jbogaardt over 4 years ago
TailBondy
now works with multiple (4D) trianglesTailBondy
computes correctly when earliest_age
is selectedfit_transform
for all tail estimators now correctly propagate all estimator attributesBondy
decay now uses the generalized Bondy formula instead of exponential decaytail_
attribute representing the point estimate of the tailattachment_age
parameter to allow for attachment before the end of the triangleTailCurve
now has slope_
and intercept_
attributes for a diagnostics of the estimator.TailBondy
now has earliest_ldf_
attributes to allow for diagnostics of the estimator.Published by jbogaardt over 4 years ago
Triangle.calendar_correlation
method and companion class CalendarCorrelation
to support detecting calendar year correlations in triangles.Triangle.developmen_correlation
method and companion class DevelopmentCorrelation
to support detecting development correlations in triangles.