- added relative cluster path in xgboost

- added MAD_Outlier_Remove Function
- added MAD_Outlier_Removal to xgboost in extra file

model_training/MAD outlier removal/mad_outlier_removal.ipynb

@@ -1,12 +1,40 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "e790b157",
+   "metadata": {},
+   "source": [
+    "Im folgenden wird auf die Daten das MAD Outlier removal angewendet."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "f38311af",
+   "id": "4bd7c061",
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from sklearn.preprocessing import StandardScaler, MinMaxScaler\n",
+    "\n",
+    "def mad_outlier_removal(df, columns, threshold=3.5):\n",
+    "    \"\"\"\n",
+    "    Entfernt Ausreißer basierend auf Median Absolute Deviation (MAD).\n",
+    "    threshold: typischer Wert ist 3.5 (entspricht robustem Z-Score Cutoff).\n",
+    "    \"\"\"\n",
+    "    df_clean = df.copy()\n",
+    "    for col in columns:\n",
+    "        median = df_clean[col].median()\n",
+    "        mad = np.median(np.abs(df_clean[col] - median))\n",
+    "        if mad == 0:\n",
+    "            continue  # keine Streuung, keine Ausreißer\n",
+    "        robust_z = 0.6745 * (df_clean[col] - median) / mad\n",
+    "        mask = np.abs(robust_z) <= threshold\n",
+    "        df_clean = df_clean[mask]\n",
+    "    return df_clean"
+   ]
   }
  ],
  "metadata": {