Update tutorial

Author: peanutfun

doc/user-guide/climada_util_calibrate.ipynb

@@ -96,7 +96,6 @@
    ],
    "source": [
     "import pandas as pd\n",
-    "from sklearn.metrics import mean_squared_log_error\n",
     "\n",
     "from climada.util import log_level\n",
     "from climada.util.api_client import Client\n",
@@ -106,6 +105,7 @@
     "    BayesianOptimizer,\n",
     "    BayesianOptimizerController,\n",
     "    OutputEvaluator,\n",
+    "    msle,\n",
     ")\n",
     "\n",
     "# Load hazard and exposure from Data API\n",
@@ -133,7 +133,7 @@
     "    # Estimated parameter bounds\n",
     "    bounds={\"v_half\": (26, 100)},\n",
     "    # Cost function\n",
-    "    cost_func=mean_squared_log_error,\n",
+    "    cost_func=msle,\n",
     "    # Transform impact to pandas Dataframe with same structure as data\n",
     "    impact_to_dataframe=lambda impact: impact.impact_at_reg(exposure.gdf[\"region_id\"]),\n",
     ")\n",
@@ -2192,7 +2192,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
+   "execution_count": null,
    "metadata": {},
    "outputs": [
     {
@@ -2211,8 +2211,12 @@
     }
    ],
    "source": [
-    "from climada.util.calibrate import Input, BayesianOptimizer, BayesianOptimizerController\n",
-    "from sklearn.metrics import mean_squared_log_error\n",
+    "from climada.util.calibrate import (\n",
+    "    Input,\n",
+    "    BayesianOptimizer,\n",
+    "    BayesianOptimizerController,\n",
+    "    msle,\n",
+    ")\n",
     "\n",
     "from climada.util import log_level\n",
     "\n",
@@ -2223,7 +2227,7 @@
     "        exposure=exposure,\n",
     "        data=data,\n",
     "        impact_func_creator=impact_func_tc,\n",
-    "        cost_func=lambda x, y, w: mean_squared_log_error(x, y, sample_weight=w),\n",
+    "        cost_func=msle,\n",
     "        impact_to_dataframe=lambda imp: imp.impact_at_reg(exposure.gdf[\"region_id\"]),\n",
     "        bounds=bounds,\n",
     "    )\n",
@@ -3726,9 +3730,8 @@
     "    Input,\n",
     "    BayesianOptimizer,\n",
     "    OutputEvaluator,\n",
-    "    BayesianOptimizerOutputEvaluator,\n",
+    "    msle,\n",
     ")\n",
-    "from sklearn.metrics import mean_squared_log_error\n",
     "import matplotlib.pyplot as plt\n",
     "\n",
     "\n",
@@ -3740,7 +3743,7 @@
     "        exposure=exposure,\n",
     "        data=data,\n",
     "        impact_func_creator=impact_func_tc,\n",
-    "        cost_func=mean_squared_log_error,\n",
+    "        cost_func=msle,\n",
     "        impact_to_dataframe=lambda imp: imp.impact_at_reg(exposure.gdf[\"region_id\"]),\n",
     "        bounds=bounds,\n",
     "        **input_kwargs,\n",
@@ -3988,7 +3991,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 35,
+   "execution_count": null,
    "metadata": {},
    "outputs": [
     {
@@ -4110,21 +4113,18 @@
     "    AverageEnsembleOptimizer,\n",
     "    BayesianOptimizer,\n",
     "    BayesianOptimizerController,\n",
+    "    msle,\n",
     ")\n",
     "from climada.util import log_level\n",
     "\n",
-    "from sklearn.metrics import mean_squared_log_error\n",
-    "\n",
     "# Define calibration input\n",
     "with log_level(\"WARNING\", name_prefix=\"climada.util.calibrate\"):\n",
     "    input = Input(\n",
     "        hazard=hazard,\n",
     "        exposure=exposure,\n",
     "        data=data,\n",
     "        impact_func_creator=impact_func_tc,\n",
-    "        cost_func=lambda data, predicted, weights: mean_squared_log_error(\n",
-    "            data, predicted, sample_weight=weights\n",
-    "        ),\n",
+    "        cost_func=msle,\n",
     "        impact_to_dataframe=lambda imp: imp.impact_at_reg(exposure.gdf[\"region_id\"]),\n",
     "        bounds=bounds,\n",
     "    )\n",
@@ -4475,7 +4475,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 34,
+   "execution_count": null,
    "metadata": {},
    "outputs": [
     {
@@ -4668,7 +4668,7 @@
     "        exposure=exposure,\n",
     "        data=data,\n",
     "        impact_func_creator=impact_func_tc,\n",
-    "        cost_func=lambda x, y, w: mean_squared_log_error(x, y, sample_weight=w),\n",
+    "        cost_func=msle,\n",
     "        impact_to_dataframe=lambda imp: imp.impact_at_reg(exposure.gdf[\"region_id\"]),\n",
     "        bounds=bounds,\n",
     "    )\n",