# version 0.2.6

• Enhancements:
  • Replaced snow with parallel. Thanks to Henrik Bengtsson. #12
  • Replaced doSNOW and foreach with pbapply for parallel processing and progress bars in calibration_glm().

Bug fixed: - calibration_glm() no longer fails when only one model is evaluated (e.g., formula_mode = "complex") with verbose = TRUE. - calibration_glm() now closes the parallel cluster even if an error occurs during model evaluation.

# version 0.2.5

Bug fixed: - Fixed: Bug in response curves with fitted models with one variable.

# version 0.2.4

• Bug fixed:
  • Fixed: Replace Rf_error with Rcpp::stop (Calls to Rf_error #11).

# version 0.2.3

• Bug fixed:
  • Fixed bug in predict_selected related to the ‘AICw’ rounded values.

# version 0.2.2

• Functions update:
  • add flag “show_lines” (Default = TRUE) in response_curve() function.

# version 0.2.1

• Bug Fixes:
  • Fixed bug in plot_niche_signal related to the ‘lwd’ argument handling.
  • Fixed bug in plot_importance that caused an error when the model list contained only one model.

# version 0.2.0

• Bug Fixes:

  • var_importance() now generates a bar plot even when the model list contains a single model, instead of throwing an error.

  • get_formulas() now returns the correct count of generated formulas when mode = "intensive".

# version 0.1.9

• New Functions:
  • Added three new functions: resp2var(), jackknife(), and plot_jk().
    • resp2var(): Transforms species probability data into a two-dimensional environmental space for visualization.
    • jackknife(): Evaluates the influence of each variable on the overall model using four distinct metrics: ROC-AUC, TSS, AICc, and Deviance. This function facilitates jackknife resampling to assess variable importance.
    • plot_jk(): A function to plot the results of the jackknife resampling.
• Bug Fixes:
  • Fixed a bug in calibration_glm() related to runtime calculation errors.

# version 0.1.8

• New Classes:
  • Added two new classes: enmpa_calibration and enmpa_fitted_models.
    • These classes help manage the list outputs from the functions calibration_glm and fit_selected.
    • Each class has two associated methods: summary() and print(), which provide summaries and print representations of the objects, respectively.
• Updates to predict_glm:
  • Added a new flag extrapolation_type to indicate the type of extrapolation:
    • "E": Free extrapolation
    • "NE": No extrapolation
    • "EC": Extrapolation with clamping
  • The flag var_to_clamp was replaced by restricted_vars.
  • The flag clamping was removed.
• Updates to model_validation:
  • Now includes ‘residual deviance’ as a validation metric.

# version 0.1.5

• Initial CRAN submission.