GP fitness function fix

Candidate expressions which give badly typed outputs are now assigned infinite fitness to signify their incorrect behaviour.

Fixed GP estimation fitness function

What's Changed

• Improved GP fitness to use NRMSE by @jmafoster1 in #300

Full Changelog: v9.0.0...v9.0.1

Added Ignore Cycles and Dropped Python 3.9

What's Changed

• Ignore cycles by @jmafoster1 in #294
• Python Version Updates to CI/CD Workflows by @f-allian in #297
• Fix: Python Version Updates to CI/CD Workflows by @f-allian in #298

Full Changelog: v8.1.0...v9.0.0

Total time in IPCW estimator

What's Changed

• Support for treatment strategies with interventions at non-consecutiv… by @jmafoster1 in #292

Full Changelog: v8.0.0...v8.1.0

GP regression equation fitting

This release provides functionality to fit regression equations using Genetic Programming. To facilitate this, we move estimation from estimation.py within the testing package to its own dedicated estimation package with each estimator being in its own separate file.

Fixed IPCW alpha bug

The value of alpha from IPCWEstimator was not being passed into the Cox fitter, so it was always using 95% confidence intervals. Now alpha is being passed in, so confidence intervals can be customised as intended.

G-methods estimation

A new estimator class IPCWEstimator for time to event analysis.

v7.0.0

From this version onwards, the CTF supports python 3.9 to 3.12

What's Changed

• Documentation Revamp by @f-allian in #279
• Python 312 compatible by @christopher-wild in #280

Full Changelog: v6.0.2...v7.0.0

v6.0.2

What's Changed

• Fix: Small changes to the automate-dafni workflow by @f-allian in #272
• Added in the control data by @jmafoster1 in #273
• Updated workflow schematic by @jmafoster1 in #275
• Fix Schematic Images by @f-allian in #278

Full Changelog: v6.0.1...v6.0.2

v6.0.1: Merge pull request #269 from CITCOM-project/interaction-terms

Temporary workaround for "I(...) not in df" bug