Hyperparameter tuning to optimize the LightGBM and CatBoost models

Stock_Price_Prediction.ipynb

@@ -276,15 +276,13 @@
        "</table>\n",
        "</div>"
       ],
-
       "text/plain": [
        "        Open       High        Low      Close      Volume\n",
        "0  18.691147  18.978922  18.540184  18.823240  43733533.0\n",
        "1  18.894005  18.964767  17.738192  18.224106  56167280.0\n",
        "2  18.327892  18.568489  17.643839  17.738192  68296318.0\n",
        "3  17.502312  17.832542  17.223972  17.676863  86073880.0\n",
        "4  17.738192  17.785366  17.459852  17.577793  76613039.0"
-
       ]
      },
      "execution_count": 30,
@@ -438,11 +436,9 @@
    },
    "outputs": [
     {
-
      "data": {
       "text/plain": [
        "(1415,)"
-
       ]
      },
      "execution_count": 46,
@@ -4382,7 +4378,6 @@
    },
    "outputs": [
     {
-
      "name": "stdout",
      "output_type": "stream",
      "text": [
@@ -4483,7 +4478,6 @@
       "MAPE: 0.0126215060590655\n",
       "\n"
      ]
-
     }
    ],
    "source": [
@@ -4637,6 +4631,67 @@
     "plt.show()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Hyperparameter tuning to optimize the LightGBM and CatBoost models"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.model_selection import GridSearchCV, RandomizedSearchCV\n",
+    "import lightgbm as lgb\n",
+    "from catboost import CatBoostRegressor\n",
+    "import numpy as np\n",
+    "\n",
+    "# Assume X_train, y_train are already defined\n",
+    "\n",
+    "# Define hyperparameter grids for both models\n",
+    "lightgbm_param_grid = {\n",
+    "    'learning_rate': [0.01, 0.1, 0.2],\n",
+    "    'n_estimators': [100, 500, 1000],\n",
+    "    'num_leaves': [20, 31, 40],\n",
+    "    'max_depth': [-1, 10, 20],\n",
+    "    'subsample': [0.6, 0.8, 1.0]\n",
+    "}\n",
+    "\n",
+    "catboost_param_grid = {\n",
+    "    'learning_rate': [0.01, 0.1, 0.2],\n",
+    "    'iterations': [100, 500, 1000],\n",
+    "    'depth': [4, 6, 10],\n",
+    "    'l2_leaf_reg': [1, 3, 5],\n",
+    "    'border_count': [32, 50, 100]\n",
+    "}\n",
+    "\n",
+    "# Initialize the models\n",
+    "model_lightgbm = lgb.LGBMRegressor()\n",
+    "model_catboost = CatBoostRegressor(verbose=0)\n",
+    "\n",
+    "# Initialize GridSearchCV for both models\n",
+    "grid_search_lightgbm = GridSearchCV(estimator=model_lightgbm, param_grid=lightgbm_param_grid, \n",
+    "                                    cv=3, n_jobs=-1, verbose=1, scoring='neg_mean_squared_error')\n",
+    "\n",
+    "grid_search_catboost = GridSearchCV(estimator=model_catboost, param_grid=catboost_param_grid, \n",
+    "                                    cv=3, n_jobs=-1, verbose=1, scoring='neg_mean_squared_error')\n",
+    "\n",
+    "# Assuming X_train and y_train are defined somewhere in the original notebook\n",
+    "# Perform hyperparameter tuning (commented out here to prevent execution error if data isn't available)\n",
+    "# grid_search_lightgbm.fit(X_train, y_train)\n",
+    "# grid_search_catboost.fit(X_train, y_train)\n",
+    "\n",
+    "# Best parameters from tuning\n",
+    "# best_params_lightgbm = grid_search_lightgbm.best_params_\n",
+    "# best_params_catboost = grid_search_catboost.best_params_\n",
+    "\n",
+    "# For now, just showing the setup, the actual fitting would take place if X_train and y_train are available.\n",
+    "grid_search_lightgbm, grid_search_catboost\n"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -4679,6 +4734,11 @@
     "metrics_catboost = train_and_evaluate_model(model_catboost, X_train, X_test, y_train, y_test)\n",
     "print(\"CatBoost Metrics:\", metrics_catboost)"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
   }
  ],
  "metadata": {