@@ -97,9 +97,33 @@ Installation
97
97
Version History
98
98
-----------------
99
99
100
- 0.18.0
100
+ 0.20.0 (release candidate for version 1.0)
101
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
102
+
103
+ * ``sev_attachment ``: changed default to ``None ``; in that case gross losses equal
104
+ ground-up losses, with no adjustment. But if layer is 10 xs 0 then losses
105
+ become conditional on X > 0. That results in a different behaviour, e.g.,
106
+ when using ``dsev[0:3] ``. Ripple through effect in Aggregate (change default),
107
+ Severity (change default, and change moment calculation; need to track the "attachment"
108
+ of zero and the fact that it came from None, to track Pr attaching)
109
+ * dsev: check if any elements are < 0 and set to zero before computing moments
110
+ in dhistogram
111
+ * same for dfreq; implemented in ``validate_discrete_distribution `` in distributions module
112
+ * Default ``recommend_p=0.99999 `` set in constsants module.
113
+ * ``interpreter_test_suite `` renamed to ``run_test_suite `` and includes test
114
+ to count and report if there are errors.
115
+ * Reason codes for failing validation; Aggregate.qt becomes Aggregte.explain_validation
116
+
117
+ 0.19.0
101
118
~~~~~~~
102
119
120
+ * Fixed reinsurance description formatting
121
+ * Improved splice parsing to allow explicit entry of lb and ub; needed to
122
+ model mixtures of mixtures (Albrecher et al. 2017)
123
+
124
+ 0.18.0 (major update)
125
+ ~~~~~~~~~~~~~~~~~~~~~~~
126
+
103
127
* Added ability to specify occ reinsurance after a built in agg; this
104
128
allows you to alter a gross aggregate more easily.
105
129
* ``Underwriter.safe_lookup `` uses deepcopy rather than copy to avoid
@@ -152,7 +176,7 @@ Version History
152
176
new arguments ``sev_lb `` and ``sev_ub ``, each lists.
153
177
* ``Underwriter.build `` defaults update argument to None, which uses the object default.
154
178
* pretty printing: now returns a value, no tacit mode; added _html version to
155
- run through pygments, that looks good in Jupyter Lab.
179
+ run through pygments, that looks good in Jupyter Lab.
156
180
157
181
0.17.1
158
182
~~~~~~~~
0 commit comments