Unreleased

iai 1.7.0

• Add cleanup_installation
• Deprecate impute and impute_cv
• Add predict_reward
• Use official julialang.org binaries in install_julia
• Support additional args in variable_importance
• Add get_machine_id
• Add glmnetcv_classifier
• Add direct score calculation
• Add tree refitting functions

iai 1.6.0

• Add add_julia_processes
• Add tree stability analysis functions
• Add get_features_used
• Add support for IAI v2.2 RewardEstimation interface
• Add predict_shap
• Add plot/autoplot functions for roc_curve and grid_search

iai 1.5.0

• Add install_julia and install_system_image for automatic installation
• Add get_grid_result_details
• Rename get_grid_results to get_grid_result_summary

iai 1.4.0

• Add support for IAI v2.1 RewardEstimation interface
• Add support for Heuristics
• Add get_roc_curve_data
• Add predict_treatment_rank and predict_treatment_outcome
• Add get_policy_treatment_outcome, get_survival_expected_time, and get_survival_hazard
• Add write_pdf and write_svg to output trees as static images

iai 1.3.0

• Add support for Optimal Policy Trees
• Add support for reward_estimator for prescription problems
• Add predict_expected_survival_time for survival learners
• Add direct roc_curve calculation from probabilities and labels
• optimal_tree_survivor has been deprecated and renamed to optimal_tree_survival_learner to match the IAI Julia interface.

iai 1.2.0

• Allow specifying Julia location with IAI_JULIA environment variable
• Fixes to ensure iai_setup is run before any Julia functions are called
• Support for CategoricalArrays.jl v0.7.7
• Add predict_hazard for survival learners

iai 1.1.0

• Add API for Optimal Feature Selection and advanced tree visualization
• Display visualizations in the viewer if available
• Update titles of documentation pages
• No longer need to run iai_setup() before using iai functions

iai 1.0.0

• Initial release