(fix) Future attributes

Some attributes were adapted so they dont give info about future, such as driverWins, constructorWins, driverPosI, etc.
pabramber01 · Apr 2, 2024 · f00c713 · f00c713
1 parent edb2b9a
commit f00c713
Show file tree

Hide file tree

Showing 42 changed files with 73,868 additions and 73,759 deletions.
diff --git a/f1_prediction/adding_data/model_validation_add.ipynb b/f1_prediction/adding_data/model_validation_add.ipynb
@@ -29,6 +29,7 @@
    "source": [
     "from sklearn.preprocessing import LabelEncoder, RobustScaler\n",
     "\n",
+    "import numpy as np\n",
     "import pandas as pd\n",
     "\n",
     "pd.set_option(\"display.max_columns\", None)\n",
@@ -119,7 +120,20 @@
     "    lambda x: x[\"driverRef\"] in x[\"driverRefPodium\"], axis=1\n",
     ")\n",
     "df[\"driverPodiums\"] = df.groupby([\"raceYear\", \"driverRef\"])[\"driverIsPodium\"].cumsum()\n",
-    "df = df.drop([\"driverRefPodium\", \"driverIsPodium\"], axis=1)"
+    "df = df.drop([\"driverRefPodium\", \"driverIsPodium\"], axis=1)\n",
+    "\n",
+    "years = df[\"raceYear\"].drop_duplicates().to_numpy()\n",
+    "drivers = df[\"driverRef\"].drop_duplicates().to_numpy()\n",
+    "\n",
+    "for year in years:\n",
+    "    for driver in drivers:\n",
+    "        mask = (df[\"raceYear\"] == year) & (df[\"driverRef\"] == driver)\n",
+    "        races = df.loc[mask, \"driverPodiums\"].iloc[:-1]\n",
+    "        races.loc[-1] = 0\n",
+    "        races.index += 1\n",
+    "        races.sort_index(inplace=True)\n",
+    "        races = races.to_numpy()\n",
+    "        df.loc[mask, \"driverPodiums\"] = races"
    ]
   },
   {
@@ -143,7 +157,40 @@
     "    df[f\"driverPos{i}\"] = df.groupby([\"raceYear\", \"driverRef\"])[\n",
     "        f\"driverIsPos{i}\"\n",
     "    ].cumsum()\n",
-    "    df = df.drop([f\"driverRefPos{i}\", f\"driverIsPos{i}\"], axis=1)"
+    "    df = df.drop([f\"driverRefPos{i}\", f\"driverIsPos{i}\"], axis=1)\n",
+    "\n",
+    "years = df[\"raceYear\"].drop_duplicates().to_numpy()\n",
+    "drivers = df[\"driverRef\"].drop_duplicates().to_numpy()\n",
+    "\n",
+    "features = [\n",
+    "    \"driverPos2\",\n",
+    "    \"driverPos3\",\n",
+    "    \"driverPos4\",\n",
+    "    \"driverPos5\",\n",
+    "    \"driverPos6\",\n",
+    "    \"driverPos7\",\n",
+    "    \"driverPos8\",\n",
+    "    \"driverPos9\",\n",
+    "    \"driverPos10\",\n",
+    "    \"driverPos11\",\n",
+    "    \"driverPos12\",\n",
+    "    \"driverPos13\",\n",
+    "    \"driverPos14\",\n",
+    "    \"driverPos15\",\n",
+    "    \"driverPos16\",\n",
+    "    \"driverPos17\",\n",
+    "    \"driverPos18\",\n",
+    "]\n",
+    "\n",
+    "for year in years:\n",
+    "    for driver in drivers:\n",
+    "        mask = (df[\"raceYear\"] == year) & (df[\"driverRef\"] == driver)\n",
+    "        races = df.loc[mask, features].iloc[:-1]\n",
+    "        races.loc[-1] = np.zeros(17, dtype=int)\n",
+    "        races.index += 1\n",
+    "        races.sort_index(inplace=True)\n",
+    "        races = races.to_numpy()\n",
+    "        df.loc[mask, features] = races"
    ]
   },
   {

diff --git a/f1_prediction/adding_data/model_validation_add_pdm.ipynb b/f1_prediction/adding_data/model_validation_add_pdm.ipynb
diff --git a/f1_prediction/adding_data/model_validation_add_posvar.ipynb b/f1_prediction/adding_data/model_validation_add_posvar.ipynb
diff --git a/f1_prediction/adding_data/model_validation_all_pdm.ipynb b/f1_prediction/adding_data/model_validation_all_pdm.ipynb
diff --git a/f1_prediction/adding_data/model_validation_all_posvar.ipynb b/f1_prediction/adding_data/model_validation_all_posvar.ipynb
diff --git a/f1_prediction/adding_data/model_validation_cir_pdm.ipynb b/f1_prediction/adding_data/model_validation_cir_pdm.ipynb
diff --git a/f1_prediction/adding_data/model_validation_cir_posvar.ipynb b/f1_prediction/adding_data/model_validation_cir_posvar.ipynb
diff --git a/f1_prediction/adding_data/model_validation_drv.ipynb b/f1_prediction/adding_data/model_validation_drv.ipynb
@@ -342,7 +342,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Encoding and normalization"
+    "## Encoding and normalization\n"
    ]
   },
   {

diff --git a/f1_prediction/adding_data/model_validation_drv_pdm.ipynb b/f1_prediction/adding_data/model_validation_drv_pdm.ipynb
diff --git a/f1_prediction/adding_data/model_validation_drv_posvar.ipynb b/f1_prediction/adding_data/model_validation_drv_posvar.ipynb
diff --git a/f1_prediction/adding_data/model_validation_wea.ipynb b/f1_prediction/adding_data/model_validation_wea.ipynb
@@ -211,7 +211,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Encoding and normalization"
+    "## Encoding and normalization\n"
    ]
   },
   {

diff --git a/f1_prediction/adding_data/model_validation_wea_pdm.ipynb b/f1_prediction/adding_data/model_validation_wea_pdm.ipynb
diff --git a/f1_prediction/adding_data/model_validation_wea_posvar.ipynb b/f1_prediction/adding_data/model_validation_wea_posvar.ipynb
diff --git a/f1_prediction/assets/data/processed/adding_data.csv b/f1_prediction/assets/data/processed/adding_data.csv
diff --git a/f1_prediction/assets/data/processed/adding_data_X.csv b/f1_prediction/assets/data/processed/adding_data_X.csv
diff --git a/f1_prediction/assets/data/processed/additional.csv b/f1_prediction/assets/data/processed/additional.csv
diff --git a/f1_prediction/assets/data/processed/additional_X.csv b/f1_prediction/assets/data/processed/additional_X.csv
diff --git a/f1_prediction/assets/data/processed/base_model.csv b/f1_prediction/assets/data/processed/base_model.csv
diff --git a/f1_prediction/assets/data/processed/base_model_X.csv b/f1_prediction/assets/data/processed/base_model_X.csv
diff --git a/f1_prediction/assets/data/processed/circuit.csv b/f1_prediction/assets/data/processed/circuit.csv
diff --git a/f1_prediction/assets/data/processed/circuit_X.csv b/f1_prediction/assets/data/processed/circuit_X.csv
diff --git a/f1_prediction/assets/data/processed/driver_ratings.csv b/f1_prediction/assets/data/processed/driver_ratings.csv
diff --git a/f1_prediction/assets/data/processed/driver_ratings_X.csv b/f1_prediction/assets/data/processed/driver_ratings_X.csv
diff --git a/f1_prediction/assets/data/processed/missing_values.csv b/f1_prediction/assets/data/processed/missing_values.csv
diff --git a/f1_prediction/assets/data/processed/weather.csv b/f1_prediction/assets/data/processed/weather.csv
diff --git a/f1_prediction/assets/data/processed/weather_X.csv b/f1_prediction/assets/data/processed/weather_X.csv
diff --git a/f1_prediction/base_model/advanced_tuning.ipynb b/f1_prediction/base_model/advanced_tuning.ipynb
@@ -4,14 +4,14 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Advanced hyperparamenter tuning"
+    "# Advanced hyperparamenter tuning\n"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Tuning methods"
+    "## Tuning methods\n"
    ]
   },
   {

diff --git a/f1_prediction/base_model/advanced_tuning_pdm.ipynb b/f1_prediction/base_model/advanced_tuning_pdm.ipynb
diff --git a/f1_prediction/base_model/advanced_tuning_posvar.ipynb b/f1_prediction/base_model/advanced_tuning_posvar.ipynb
diff --git a/f1_prediction/base_model/feature_selection_pdm.ipynb b/f1_prediction/base_model/feature_selection_pdm.ipynb
diff --git a/f1_prediction/base_model/feature_selection_posvar.ipynb b/f1_prediction/base_model/feature_selection_posvar.ipynb
diff --git a/f1_prediction/base_model/model_validation.ipynb b/f1_prediction/base_model/model_validation.ipynb
diff --git a/f1_prediction/base_model/model_validation_pdm.ipynb b/f1_prediction/base_model/model_validation_pdm.ipynb
diff --git a/f1_prediction/base_model/model_validation_pos.ipynb b/f1_prediction/base_model/model_validation_pos.ipynb
diff --git a/f1_prediction/base_model/model_validation_posfxd.ipynb b/f1_prediction/base_model/model_validation_posfxd.ipynb
diff --git a/f1_prediction/base_model/model_validation_posvar.ipynb b/f1_prediction/base_model/model_validation_posvar.ipynb
diff --git a/f1_prediction/base_model/model_validation_wnr.ipynb b/f1_prediction/base_model/model_validation_wnr.ipynb
diff --git a/f1_prediction/base_model/simple_tuning_pdm.ipynb b/f1_prediction/base_model/simple_tuning_pdm.ipynb
@@ -117,7 +117,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "KNeighborsClassifier: 0.726096099171753 with {'metric': 'cosine', 'n_neighbors': 5,\n",
+      "KNeighborsClassifier: 0.7215047203351547 with {'metric': 'cosine', 'n_neighbors': 5,\n",
       "\t'weights': 'distance'}\n"
      ]
     }
@@ -155,8 +155,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "DecisionTreeClassifier: 0.784020180449896 with {'criterion': 'gini', 'max_depth': 3,\n",
-      "\t'splitter': 'best'}\n"
+      "DecisionTreeClassifier: 0.7919333040678084 with {'criterion': 'log_loss', 'max_depth':\n",
+      "\t4, 'splitter': 'best'}\n"
      ]
     }
    ],
@@ -193,8 +193,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "RandomForestClassifier: 0.7735981616295589 with {'criterion': 'gini', 'max_depth': 10,\n",
-      "\t'n_estimators': 50}\n"
+      "RandomForestClassifier: 0.7647004284290003 with {'criterion': 'gini', 'max_depth': 10,\n",
+      "\t'n_estimators': 200}\n"
      ]
     }
    ],
@@ -231,7 +231,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "MLPClassifier: 0.8008305374733512 with {'activation': 'logistic', 'hidden_layer_sizes':\n",
+      "MLPClassifier: 0.797863262800811 with {'activation': 'logistic', 'hidden_layer_sizes':\n",
       "\t(50, 20, 5)}\n"
      ]
     }
@@ -265,10 +265,10 @@
    "source": [
     "After viewing several runs, the hyperparameters for each algorithm are as follows\n",
     "\n",
-    "- KNeighborsClassifier: 0.7263922435898924 with {'metric': 'cosine', 'n_neighbors': 5, 'weights': 'distance'}\n",
-    "- DecisionTreeClassifier: 0.7892538259292045 with {'criterion': 'gini', 'max_depth': 3, 'splitter': 'best'}\n",
-    "- RandomForestClassifier: 0.7778548164328474 with {'criterion': 'gini', 'max_depth': 10, 'n_estimators': 200}\n",
-    "- MLPClassifier: 0.799283305036914 with {'activation': 'logistic', 'hidden_layer_sizes': (50, 20, 5)}\n"
+    "- KNeighborsClassifier: 0.7215047203351547 with {'metric': 'cosine', 'n_neighbors': 5, 'weights': 'distance'}\n",
+    "- DecisionTreeClassifier: 0.7919333040678084 with {'criterion': 'log_loss', 'max_depth': 4, 'splitter': 'best'}\n",
+    "- RandomForestClassifier: 0.7647004284290003 with {'criterion': 'gini', 'max_depth': 10, 'n_estimators': 200}\n",
+    "- MLPClassifier: 0.797863262800811 with {'activation': 'logistic', 'hidden_layer_sizes': (50, 20, 5)}\n"
    ]
   }
  ],

diff --git a/f1_prediction/base_model/simple_tuning_posvar.ipynb b/f1_prediction/base_model/simple_tuning_posvar.ipynb
@@ -39,10 +39,7 @@
     "sys.path.append(\"..\")\n",
     "\n",
     "from utils.custom_cvs import VariableTimeSeriesSplit\n",
-    "from utils.custom_scorers import (\n",
-    "    balanced_accuracy_1interval_score,\n",
-    "    mean_absolute_1interval_error,\n",
-    ")\n",
+    "from utils.custom_scorers import balanced_accuracy_1interval_score\n",
     "\n",
     "import textwrap\n",
     "import numpy as np\n",
@@ -114,8 +111,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "KNeighborsClassifier: 0.28718245979609613 with {'metric': 'manhattan', 'n_neighbors':\n",
-      "\t101, 'weights': 'distance'}\n"
+      "KNeighborsClassifier: 0.286227602363966 with {'metric': 'manhattan', 'n_neighbors': 101,\n",
+      "\t'weights': 'uniform'}\n"
      ]
     }
    ],
@@ -152,8 +149,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "DecisionTreeClassifier: 0.3389971896790078 with {'criterion': 'entropy', 'max_depth': 5,\n",
-      "\t'splitter': 'random'}\n"
+      "DecisionTreeClassifier: 0.3423765333992606 with {'criterion': 'entropy', 'max_depth': 4,\n",
+      "\t'splitter': 'best'}\n"
      ]
     }
    ],
@@ -190,8 +187,8 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "RandomForestClassifier: 0.34932197852652397 with {'criterion': 'gini', 'max_depth': 5,\n",
-      "\t'n_estimators': 50}\n"
+      "RandomForestClassifier: 0.3478434171047807 with {'criterion': 'log_loss', 'max_depth':\n",
+      "\t5, 'n_estimators': 200}\n"
      ]
     }
    ],
@@ -228,7 +225,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "MLPClassifier: 0.35292989628216903 with {'activation': 'logistic', 'hidden_layer_sizes':\n",
+      "MLPClassifier: 0.3517574344847072 with {'activation': 'logistic', 'hidden_layer_sizes':\n",
       "\t(50, 25)}\n"
      ]
     }
@@ -262,10 +259,10 @@
    "source": [
     "After viewing several runs, the hyperparameters for each algorithm are as follows\n",
     "\n",
-    "- KNeighborsClassifier: 0.2863537094218912 with {'metric': 'manhattan', 'n_neighbors': 101, 'weights': 'uniform'}\n",
-    "- DecisionTreeClassifier: 0.34112213333804237 with {'criterion': 'gini', 'max_depth': 4, 'splitter': 'best'}\n",
-    "- RandomForestClassifier: 0.3479073809187445 with {'criterion': 'gini', 'max_depth': 5, 'n_estimators': 200}\n",
-    "- MLPClassifier: 0.3586966547762002 with {'activation': 'logistic', 'hidden_layer_sizes': (50, 25)}\n"
+    "- KNeighborsClassifier: 0.286227602363966 with {'metric': 'manhattan', 'n_neighbors': 101, 'weights': 'uniform'}\n",
+    "- DecisionTreeClassifier: 0.3413275823503096 with {'criterion': 'entropy', 'max_depth': 4, 'splitter': 'best'}\n",
+    "- RandomForestClassifier: 0.34519032104259373 with {'criterion': 'gini', 'max_depth': 5, 'n_estimators': 200}\n",
+    "- MLPClassifier: 0.3515771497021497 with {'activation': 'logistic', 'hidden_layer_sizes': (50, 25)}\n"
    ]
   }
  ],

diff --git a/f1_prediction/preprocessing/data_selection.ipynb b/f1_prediction/preprocessing/data_selection.ipynb
@@ -4,14 +4,14 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Data selection"
+    "# Data selection\n"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Dependencies"
+    "## Dependencies\n"
    ]
   },
   {
@@ -41,7 +41,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Merging data"
+    "## Merging data\n"
    ]
   },
   {
@@ -202,7 +202,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Renaming data"
+    "## Renaming data\n"
    ]
   },
   {

diff --git a/f1_prediction/preprocessing/data_transformation.ipynb b/f1_prediction/preprocessing/data_transformation.ipynb
@@ -4,14 +4,14 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Data transformation"
+    "# Data transformation\n"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Dependencies"
+    "## Dependencies\n"
    ]
   },
   {
@@ -43,7 +43,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Datatype conversion"
+    "## Datatype conversion\n"
    ]
   },
   {