set meaningful names to models (#90)

mljar · May 27, 2020 · 803b9e4 · 803b9e4
1 parent f076bf6
commit 803b9e4
Show file tree

Hide file tree

Showing 3 changed files with 13 additions and 7 deletions.
diff --git a/examples/scripts/multi_class_classifier.py b/examples/scripts/multi_class_classifier.py
@@ -15,7 +15,7 @@
 
 automl = AutoML(
     #results_path="AutoML_22",
-    algorithms=["Nearest Neighbors"],
+    #algorithms=["Nearest Neighbors"],
     #algorithms=["Random Forest"],
 
     #    "Linear",

diff --git a/supervised/automl.py b/supervised/automl.py
@@ -660,7 +660,7 @@ def stacked_ensemble_step(self):
 
             params = copy.deepcopy(m.params)
             params["validation"]["X_train_path"] = X_train_stacked_path
-            params["name"] = "stacked_" + params["name"]
+            params["name"] = "Stacked" + params["name"]
             params["is_stacked"] = True
 
             if "model_architecture_json" in params["learner"]:

diff --git a/supervised/tuner/mljar_tuner.py b/supervised/tuner/mljar_tuner.py
@@ -43,6 +43,9 @@ def __init__(
 
         self._unique_params_keys = []
 
+    def get_model_name(self, model_type, models_cnt, special = ""):
+        return special + model_type.replace(" ", "")+ f"_{models_cnt}"
+
     def simple_algorithms_params(self):
         models_cnt = 0
         generated_params = []
@@ -55,7 +58,8 @@ def simple_algorithms_params(self):
                 params = self._get_model_params(model_type, seed=i + 1)
                 if params is None:
                     continue
-                params["name"] = f"model_{models_cnt + 1}"
+
+                params["name"] = self.get_model_name(model_type, models_cnt+1)
 
                 unique_params_key = MljarTuner.get_params_key(params)
                 if unique_params_key not in self._unique_params_keys:
@@ -102,8 +106,8 @@ def default_params(self, models_cnt):
                 model_type, seed=models_cnt + 1, params_type="default"
             )
             if params is None:
-                continue
-            params["name"] = f"model_{models_cnt + 1}"
+                continue 
+            params["name"] = self.get_model_name(model_type, models_cnt+1, special="Default")
 
             unique_params_key = MljarTuner.get_params_key(params)
             if unique_params_key not in self._unique_params_keys:
@@ -139,7 +143,8 @@ def get_not_so_random_params(self, models_cnt):
                 params = self._get_model_params(model_type, seed=i + 1)
                 if params is None:
                     continue
-                params["name"] = f"model_{models_cnt + 1}"
+
+                params["name"] = self.get_model_name(model_type, models_cnt+1)
 
                 unique_params_key = MljarTuner.get_params_key(params)
                 if unique_params_key not in self._unique_params_keys:
@@ -201,7 +206,8 @@ def get_hill_climbing_params(self, current_models):
                         if p is not None:
                             all_params = copy.deepcopy(m.params)
                             all_params["learner"] = p
-                            all_params["name"] = f"model_{model_max_index + 1}"
+
+                            all_params["name"] = self.get_model_name(all_params["learner"]["model_type"], model_max_index+1)
 
                             unique_params_key = MljarTuner.get_params_key(all_params)
                             if unique_params_key not in self._unique_params_keys: