bug in univariate gridsearch

tkchafin · May 15, 2024 · 32e2ffe · 32e2ffe
1 parent 9d570b0
commit 32e2ffe
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/src/resistnet/model_optimisation.py b/src/resistnet/model_optimisation.py
@@ -953,7 +953,7 @@ def optimise_univariate(self, fitmetric="aic", threads=1,
             df = pd.DataFrame(results_data)
             for var in self.resistance_network.variables:
                 var_df = df[df['variable'] == var]
-                best = var_df.loc[var_df['fitness'] == var_df['fitness'].min()]
+                best = var_df.loc[var_df['fitness'] == var_df['fitness'].max()]
                 best = best.sort_values(by=['shape', 'transform']).iloc[0]
                 best_params[var] = {
                     'transform': best['transform'],