refit model before sensitivity analysis

Author: mastoffel

.gitignore

@@ -27,4 +27,8 @@ Thumbs.db
 *.DS_Store
 
 # VScode settings
-.vscode/
+.vscode/
+
+# Quarto
+README.html
+README_files/

README.md

@@ -22,7 +22,7 @@ There's currently a lot of development, so we recommend installing the most curr
 pip install git+https://github.com/alan-turing-institute/autoemulate.git
 ```
 
-There's also a release available on PyPI (will not contain the most recent features and models)
+There's also a release available on PyPI (note: currently and older version and out of date with the documentation)
 ```bash
 pip install autoemulate
 ```
@@ -47,19 +47,27 @@ from autoemulate.simulations.projectile import simulate_projectile
 lhd = LatinHypercube([(-5., 1.), (0., 1000.)])
 X = lhd.sample(100)
 y = np.array([simulate_projectile(x) for x in X])
+
 # compare emulator models
 ae = AutoEmulate()
 ae.setup(X, y)
-best_model = ae.compare() 
+best_emulator = ae.compare() 
+
 # training set cross-validation results
 ae.summarise_cv() 
 ae.plot_cv()
+
 # test set results for the best model
-ae.evaluate(best_model) 
-ae.plot_eval(best_model)
+ae.evaluate(best_emulator) 
+ae.plot_eval(best_emulator)
+
 # refit on full data and emulate!
-best_model = ae.refit(best_model) 
-best_model.predict(X)
+emulator = ae.refit(best_emulator) 
+emulator.predict(X)
+
+# global sensitivity analysis
+si = ae.sensitivity_analysis(emulator)
+ae.plot_sensitivity_analysis(si)
 ```
 
 ## documentation

autoemulate/compare.py

@@ -557,9 +557,9 @@ def sensitivity_analysis(
         if model is None:
             if not hasattr(self, "best_model"):
                 raise RuntimeError("Must run compare() before sensitivity_analysis()")
-            model = self.best_model
+            model = self.refit(self.best_model)
             self.logger.info(
-                f"No model provided, using {get_model_name(model)}, which had the highest average cross-validation score."
+                f"No model provided, using {get_model_name(model)}, which had the highest average cross-validation score, refitted on full data."
             )
 
         Si = sensitivity_analysis(model, problem, self.X, N, conf_level, as_df)