{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "view-in-github"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/drive/12zwwBI7Fgv_MjDARHPwCr0zUKwBJm6DR\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "0k3i4qXqJ0wo"
   },
   "source": [
    "# Workshop: Implementing basic machine and deep learning for ORN classification \n",
    "\n",
    "---\n",
    "\n",
    "\n",
    "# Harini Veeraraghavan, Ph.D.\n",
    "\n",
    "## Audience:\n",
    "Data science class: Lecture 1 and 2 concepts\n",
    "\n",
    "## Data description:\n",
    "\n",
    "Data from the paper *Clinical and dosimetric dataset of time-to-event normal tissue complication probability for osteoradionecrosis* https://www.nature.com/articles/s41597-025-06321-w \n",
    "\n",
    "This is a 1,129 patient dataset of head and neck cancer patients who were treated with radiotherapy at MD Anderson Cancer Center and assessed for osteoradionecrosis (ORN) of the mandible at multiple follow ups. \n",
    "\n",
    "Data includes the following variables (Table 1 of the paper; the CSV column name is shown for each):\n",
    "\n",
    "| Category | Variable | CSV column | Coding | Units |\n",
    "|---|---|---|---|---|\n",
    "| **Demographic** | Gender | `Gender` | Male / Female | N/A |\n",
    "| | Age | `Age` | Quantitative | years |\n",
    "| | Smoking status | `Smoking status` | Current / Former / Never smoker | N/A |\n",
    "| | Smoking pack-years | `Smoking Pack-Years` | Quantitative | pack-years |\n",
    "| **Clinical** | Overall survival | `Overall survival` | 0 = deceased at last F/U, 1 = alive at last F/U | N/A |\n",
    "| | ORNJ status (**outcome**) | `ORN status` | 0 = negative, 1 = positive for ORNJ | N/A |\n",
    "| | Time to event | `Time to event` | Quantitative | months |\n",
    "| | ORNJ grade* | `ORN Grade (Tsai)` | 1 = minimal bone exposure, conservative management<br>2 = minor debridement<br>3 = HBO needed<br>4 = major surgery | N/A |\n",
    "| | Pre-RT dental extractions | `Pre-RT dental extractions` | 0 = no, 1 = yes | N/A |\n",
    "| | Chemotherapy | `Chemotherapy` | Induction / Induction and concurrent / Concurrent / None | N/A |\n",
    "| | Postoperative vs. definitive RT | `Postop RT vs. Definitive RT` | Postoperative / Definitive | N/A |\n",
    "| | HPV/p16 status | `HPV/P16 +Ve` | Yes / No / Unknown | N/A |\n",
    "| | Tumor site group | `Tumor site group` | Oropharynx / Oral cavity / Nasopharynx, nasal cavity, paranasal sinuses / Larynx, hypopharynx / Other | N/A |\n",
    "| | Mandible volume | `Mandible volume (cc)` | Quantitative | cm³ |\n",
    "| **Dosimetric** | V5Gy – V80Gy (in 5 Gy steps) | `V5Gy` … `V80Gy` | Quantitative | % of mandible volume† |\n",
    "| | D0.5% – D99.5% | `D0.5%` … `D99.5%` | Quantitative | Gy |\n",
    "\n",
    "\\*Adapted from the Tsai staging system. F/U = follow-up; HBO = hyperbaric oxygen; Tumor site \"Other\" = other pharynx and oral cavity.\n",
    "\n",
    "*Source: Table 1, [Scientific Data (2025) s41597-025-06321-w](https://www.nature.com/articles/s41597-025-06321-w/tables/1).*\n",
    "\n",
    "## Clinical data description\n",
    "Osteoradionecrosis of the jaw (ORNJ) is a radiation-induced late toxicity that can dramatically decrease patients’ quality of life. Recent increases in survival rates of head and neck cancers associated with human papillomavirus (HPV) infection have resulted in a higher frequency of radiation-induced toxicities, particularly ORNJ.\n",
    "\n",
    "Osteoradionecrosis (ORN) of the mandible is a debilitating late toxicity following head and neck radiotherapy, often resulting in pain, functional impairment, and diminished quality of life.\n",
    "Despite decades of clinical experience, predicting ORN remains difficult due to the interplay between dose, anatomy, and patient-specific biological factors.\n",
    "\n",
    "Dose-Volume Histogram (DVH) parameters summarize the spatial dose distribution within the mandible, linking radiotherapy physics to biological outcomes. However, DVHs are inherently correlated, incomplete, and sensitive to clinical data entry biases — making them an excellent testbed for exploring how data quality and preprocessing influence ML/DL model behavior.\n",
    "\n",
    "\n",
    "### Focus:\n",
    "Hands-on exercise to visualize and explore dataset, perform statistical tests, perform logistic regression and look at the results.\n",
    "\n",
    "## Make sure to create a Data folder in your local drive or in google colab and copy the TTE_ORN_data_Figshare.csv posted in the moodle\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "aXz1Q5nBJ0wq",
    "outputId": "c5e1cc3e-b9bc-47ea-93e3-1ca9d77a5c37"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Environment setup complete.\n",
      "NumPy version: 2.3.5\n",
      "Pandas version: 2.3.3\n",
      "Scikit-learn version: 1.7.2\n"
     ]
    }
   ],
   "source": [
    "# ============================================\n",
    "# 1. Environment Setup and Dependencies\n",
    "# ============================================\n",
    "\n",
    "# This cell installs required packages, configures plotting, and ensures reproducibility for the workshop.\n",
    "import os\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from pathlib import Path\n",
    "from scipy import stats\n",
    "import warnings\n",
    "from IPython.display import display\n",
    "from sklearn.metrics import ConfusionMatrixDisplay\n",
    "\n",
    "warnings.filterwarnings(\"ignore\", message=\"Precision loss\")   # harmless: a few genes are constant within a group\n",
    "sns.set_theme(style=\"whitegrid\", context=\"notebook\")\n",
    "rng = np.random.default_rng(2026)          # use this generator for anything random\n",
    "DATA_DIR = next((p for p in [Path(\"data\"), Path(\"../data\")] if p.exists()), Path(\"data\"))\n",
    "\n",
    "def ok(msg=\"looks good\"):\n",
    "    print(\"✅\", msg)\n",
    "\n",
    "from sklearn import __version__ as sklearn_version  # Import sklearn version\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, roc_curve, auc, confusion_matrix\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.inspection import permutation_importance\n",
    "from sklearn.base import BaseEstimator, ClassifierMixin\n",
    "from sklearn.metrics import make_scorer, roc_auc_score\n",
    "import statsmodels.api as sm\n",
    "import re\n",
    "\n",
    "# Configure environment\n",
    "warnings.filterwarnings('ignore')\n",
    "np.random.seed(42)\n",
    "\n",
    "# Set visualization parameters\n",
    "plt.rcParams['figure.figsize'] = (10, 6)\n",
    "plt.rcParams['font.size'] = 10\n",
    "plt.rcParams['axes.grid'] = True\n",
    "plt.rcParams['grid.alpha'] = 0.3\n",
    "\n",
    "# Install dependencies if required\n",
    "def install_dependencies():\n",
    "    try:\n",
    "        import sklearn, imblearn, statsmodels, openpyxl\n",
    "    except ImportError:\n",
    "        print(\"Installing required packages...\")\n",
    "        os.system('pip install -q scikit-learn imbalanced-learn statsmodels torch==2.* openpyxl --extra-index-url https://download.pytorch.org/whl/cpu')\n",
    "\n",
    "install_dependencies()\n",
    "\n",
    "print(\"Environment setup complete.\")\n",
    "print(f\"NumPy version: {np.__version__}\")\n",
    "print(f\"Pandas version: {pd.__version__}\")\n",
    "print(f\"Scikit-learn version: {sklearn_version}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "JuimIjhhYaX_"
   },
   "source": [
    "In this step, you will load a dataset and check if the data looks OK.\n",
    "* Look at the data and discuss amongst yourself what entries you think are useful for performing analysis and which might not be.\n",
    "* Some of the entries that have mostly Nans or have descriptive text cannot be handled by machine learning models and need to be removed.\n",
    "* Check to see that there are no NaNs and additional entries that are irrelevant to data analysis and remove them for analysis.\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 820
    },
    "id": "7kQ2rY4TJ0wr",
    "outputId": "00ed89a9-7708-47dc-d551-e85b5eb25be8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Loading DVH ORN dataset\n",
      "(1129, 61)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Study ID</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Age</th>\n",
       "      <th>Overall survival</th>\n",
       "      <th>Overall survival time</th>\n",
       "      <th>ORN status</th>\n",
       "      <th>Time to event</th>\n",
       "      <th>ORN Grade (Tsai)</th>\n",
       "      <th>Smoking status</th>\n",
       "      <th>Smoking Pack-Years</th>\n",
       "      <th>...</th>\n",
       "      <th>D70%</th>\n",
       "      <th>D75%</th>\n",
       "      <th>D80%</th>\n",
       "      <th>D85%</th>\n",
       "      <th>D90%</th>\n",
       "      <th>D95%</th>\n",
       "      <th>D97%</th>\n",
       "      <th>D98%</th>\n",
       "      <th>D99%</th>\n",
       "      <th>D99.5%</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>Male</td>\n",
       "      <td>58</td>\n",
       "      <td>0</td>\n",
       "      <td>117.0</td>\n",
       "      <td>0</td>\n",
       "      <td>117.40</td>\n",
       "      <td>0</td>\n",
       "      <td>Never</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>18.73</td>\n",
       "      <td>17.27</td>\n",
       "      <td>15.90</td>\n",
       "      <td>12.85</td>\n",
       "      <td>10.42</td>\n",
       "      <td>8.29</td>\n",
       "      <td>7.52</td>\n",
       "      <td>7.02</td>\n",
       "      <td>6.53</td>\n",
       "      <td>6.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4</td>\n",
       "      <td>Male</td>\n",
       "      <td>58</td>\n",
       "      <td>1</td>\n",
       "      <td>45.6</td>\n",
       "      <td>0</td>\n",
       "      <td>45.40</td>\n",
       "      <td>0</td>\n",
       "      <td>Never</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>45.09</td>\n",
       "      <td>42.90</td>\n",
       "      <td>39.74</td>\n",
       "      <td>29.65</td>\n",
       "      <td>14.91</td>\n",
       "      <td>7.18</td>\n",
       "      <td>5.93</td>\n",
       "      <td>5.23</td>\n",
       "      <td>3.72</td>\n",
       "      <td>2.59</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>Male</td>\n",
       "      <td>82</td>\n",
       "      <td>1</td>\n",
       "      <td>134.2</td>\n",
       "      <td>0</td>\n",
       "      <td>117.50</td>\n",
       "      <td>0</td>\n",
       "      <td>Former</td>\n",
       "      <td>70.0</td>\n",
       "      <td>...</td>\n",
       "      <td>45.12</td>\n",
       "      <td>42.07</td>\n",
       "      <td>39.17</td>\n",
       "      <td>33.16</td>\n",
       "      <td>25.49</td>\n",
       "      <td>12.51</td>\n",
       "      <td>10.62</td>\n",
       "      <td>9.80</td>\n",
       "      <td>8.18</td>\n",
       "      <td>6.42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>Male</td>\n",
       "      <td>70</td>\n",
       "      <td>1</td>\n",
       "      <td>46.2</td>\n",
       "      <td>0</td>\n",
       "      <td>46.00</td>\n",
       "      <td>0</td>\n",
       "      <td>Former</td>\n",
       "      <td>35.0</td>\n",
       "      <td>...</td>\n",
       "      <td>28.64</td>\n",
       "      <td>26.87</td>\n",
       "      <td>23.97</td>\n",
       "      <td>15.33</td>\n",
       "      <td>10.40</td>\n",
       "      <td>7.00</td>\n",
       "      <td>5.57</td>\n",
       "      <td>4.60</td>\n",
       "      <td>3.67</td>\n",
       "      <td>3.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>Female</td>\n",
       "      <td>66</td>\n",
       "      <td>0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>1</td>\n",
       "      <td>5.45</td>\n",
       "      <td>4</td>\n",
       "      <td>Never</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>48.14</td>\n",
       "      <td>28.28</td>\n",
       "      <td>10.60</td>\n",
       "      <td>4.31</td>\n",
       "      <td>2.01</td>\n",
       "      <td>1.19</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.88</td>\n",
       "      <td>0.81</td>\n",
       "      <td>0.78</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 61 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Study ID  Gender  Age  Overall survival  Overall survival time  ORN status  \\\n",
       "0         2    Male   58                 0                  117.0           0   \n",
       "1         4    Male   58                 1                   45.6           0   \n",
       "2         8    Male   82                 1                  134.2           0   \n",
       "3         9    Male   70                 1                   46.2           0   \n",
       "4        15  Female   66                 0                   86.0           1   \n",
       "\n",
       "   Time to event  ORN Grade (Tsai) Smoking status  Smoking Pack-Years  ...  \\\n",
       "0         117.40                 0          Never                 0.0  ...   \n",
       "1          45.40                 0          Never                 0.0  ...   \n",
       "2         117.50                 0         Former                70.0  ...   \n",
       "3          46.00                 0         Former                35.0  ...   \n",
       "4           5.45                 4          Never                 0.0  ...   \n",
       "\n",
       "    D70%   D75%   D80%   D85%   D90%   D95%   D97%  D98%  D99%  D99.5%  \n",
       "0  18.73  17.27  15.90  12.85  10.42   8.29   7.52  7.02  6.53    6.30  \n",
       "1  45.09  42.90  39.74  29.65  14.91   7.18   5.93  5.23  3.72    2.59  \n",
       "2  45.12  42.07  39.17  33.16  25.49  12.51  10.62  9.80  8.18    6.42  \n",
       "3  28.64  26.87  23.97  15.33  10.40   7.00   5.57  4.60  3.67    3.03  \n",
       "4  48.14  28.28  10.60   4.31   2.01   1.19   0.98  0.88  0.81    0.78  \n",
       "\n",
       "[5 rows x 61 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Study ID', 'Gender', 'Age', 'Overall survival', 'Overall survival time', 'ORN', 'Time to event', 'ORN Grade (Tsai)', 'Smoking', 'Smoking-years', 'Extractions', 'T Stage', 'N Stage', 'Chemotherapy', 'PostRT', 'HPV', 'Site', 'Mandible Volume', 'V5Gy', 'V10Gy', 'V15Gy', 'V20Gy', 'V25Gy', 'V30Gy', 'V35Gy', 'V40Gy', 'V45Gy', 'V50Gy', 'V55Gy', 'V60Gy', 'V65Gy', 'V70Gy', 'V75Gy', 'V80Gy', 'D0.5%', 'D1%', 'D2%', 'D3%', 'D5%', 'D10%', 'D15%', 'D20%', 'D25%', 'D30%', 'D35%', 'D40%', 'D45%', 'D50%', 'D55%', 'D60%', 'D65%', 'D70%', 'D75%', 'D80%', 'D85%', 'D90%', 'D95%', 'D97%', 'D98%', 'D99%', 'D99.5%']\n"
     ]
    }
   ],
   "source": [
    "# ============================================\n",
    "# 2. Step 1 — Load and Inspect the Dataset\n",
    "# ============================================\n",
    "\n",
    "print(\"\\nLoading DVH ORN dataset\")\n",
    "\n",
    "import os\n",
    "\n",
    "path = os.path.join(DATA_DIR, \"TTE_ORN_data_Figshare.csv\") \n",
    "ORN_data = pd.read_csv(path)\n",
    "\n",
    "print(ORN_data.shape)\n",
    "display(ORN_data.head())\n",
    "\n",
    "df = ORN_data\n",
    "df = df.replace(\"Unknown\", \"na\")\n",
    "df = df.rename(columns={'Tumor site group': 'Site','Mandible volume (cc)': 'Mandible Volume', 'HPV/P16 +Ve': 'HPV', 'ORN status': 'ORN', 'Postop RT vs. Definitive RT': 'PostRT', 'Smoking status': 'Smoking', 'Smoking Pack-Years': 'Smoking-years', 'Pre-RT dental extractions': 'Extractions'})\n",
    "\n",
    "## Remove entries that have a lot of Nans and are not relevant for analysis\n",
    "df = df.dropna()\n",
    "\n",
    "### print column names\n",
    "print(df.columns.tolist())\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use the column names to update the features to be used for building the classifier. \n",
    "* Hint: ORN is the class we want to predict\n",
    "* Based on teh paper by Lisanne van Dijk et.al https://pubmed.ncbi.nlm.nih.gov/33965514/, the D30% and V50Gy showed the best classification performance with AUCs of 0.76, Chemotherapy, Dental Extraction or Extractions, Post-RT (PostRT) were associated with outcomes. In the feature list, use these features, as well as Mandible Volume, Age, Gender, Smoking, Smoking-years, Extracts, HPV, ,gender, and site."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 761
    },
    "id": "n9rWqH3TJ0ws",
    "outputId": "d272a364-dc28-470e-8a63-a73f6668d7ca"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['D30%', 'V50Gy', 'Mandible Volume', 'Age', 'Gender', 'Site', 'Chemotherapy', 'PostRT', 'Extractions']\n",
      "Initial shape: (1129, 9)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>D30%</th>\n",
       "      <th>V50Gy</th>\n",
       "      <th>Mandible Volume</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Site</th>\n",
       "      <th>Chemotherapy</th>\n",
       "      <th>PostRT</th>\n",
       "      <th>Extractions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>55.91</td>\n",
       "      <td>35.21</td>\n",
       "      <td>113.43</td>\n",
       "      <td>58</td>\n",
       "      <td>Male</td>\n",
       "      <td>Oral Cavity</td>\n",
       "      <td>No Chemotherapy</td>\n",
       "      <td>postop_RT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>55.22</td>\n",
       "      <td>55.45</td>\n",
       "      <td>84.56</td>\n",
       "      <td>58</td>\n",
       "      <td>Male</td>\n",
       "      <td>Oral Cavity</td>\n",
       "      <td>Concurrent Chemotherapy</td>\n",
       "      <td>postop_RT</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>55.97</td>\n",
       "      <td>56.73</td>\n",
       "      <td>88.24</td>\n",
       "      <td>82</td>\n",
       "      <td>Male</td>\n",
       "      <td>Oral Cavity</td>\n",
       "      <td>No Chemotherapy</td>\n",
       "      <td>postop_RT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    D30%  V50Gy  Mandible Volume  Age Gender         Site  \\\n",
       "0  55.91  35.21           113.43   58   Male  Oral Cavity   \n",
       "1  55.22  55.45            84.56   58   Male  Oral Cavity   \n",
       "2  55.97  56.73            88.24   82   Male  Oral Cavity   \n",
       "\n",
       "              Chemotherapy     PostRT  Extractions  \n",
       "0          No Chemotherapy  postop_RT            0  \n",
       "1  Concurrent Chemotherapy  postop_RT            1  \n",
       "2          No Chemotherapy  postop_RT            0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ============================================\n",
    "# 2.1 Step 1 — Preprocessing to extract data entries for processing using Robust column mapping\n",
    "# ============================================\n",
    "\n",
    "import random\n",
    "\n",
    "# Map columns from the dataset to select specific columns that might be useful for analysis\n",
    "\n",
    "## Example mapping of output class to y \n",
    "y = df['ORN'].astype(int)\n",
    "\n",
    "## Do the same for input features\n",
    "#features = ['dose_max', 'dose_mean', 'Volume', 'HPV', 'Age', 'Sex', 'Primary Tumor Location']\n",
    "\n",
    "\n",
    "features = ['D30%', 'V50Gy', 'Mandible Volume', 'Age', 'Gender', 'Site', 'Chemotherapy', 'PostRT', 'Extractions']\n",
    "\n",
    "print(features)\n",
    "features = list(dict.fromkeys(features))  # Dedupe, keep order\n",
    "X = df[features].copy()\n",
    "\n",
    "print(f\"Initial shape: {X.shape}\")\n",
    "display(X.head(3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    D30%  V50Gy  Mandible Volume  Age  Extractions  Gender_Male  \\\n",
      "0  55.91  35.21           113.43   58            0         True   \n",
      "1  55.22  55.45            84.56   58            1         True   \n",
      "\n",
      "   Site_Major Salivary Glands  \\\n",
      "0                       False   \n",
      "1                       False   \n",
      "\n",
      "   Site_Nasopharynx, Nasal Cavity, Paranasal Sinuses  Site_Oral Cavity  \\\n",
      "0                                              False              True   \n",
      "1                                              False              True   \n",
      "\n",
      "   Site_Oropharynx  Site_Other  Chemotherapy_Induction Chemotherapy  \\\n",
      "0            False       False                                False   \n",
      "1            False       False                                False   \n",
      "\n",
      "   Chemotherapy_Induction and Concurrent  Chemotherapy_No Chemotherapy  \\\n",
      "0                                  False                          True   \n",
      "1                                  False                         False   \n",
      "\n",
      "   PostRT_postop_RT  \n",
      "0              True  \n",
      "1              True  \n"
     ]
    }
   ],
   "source": [
    "X_num = X.copy()\n",
    "y = y.copy()\n",
    "\n",
    "non_numeric_cols = X_num.select_dtypes(exclude=['number']).columns.tolist()\n",
    "\n",
    "# One-hot encode any remaining truly categorical columns\n",
    "X_num = pd.get_dummies(X_num, columns=non_numeric_cols, drop_first=True)\n",
    "display(X_num.head(2))\n",
    "\n",
    "# ---------- Make sure to change Nan values to numeric values -- ML models cannot understand non-numbers ----------\n",
    "X_filled = X_num.fillna(X_num.median(numeric_only=True))\n",
    "X_res = X_filled.copy()\n",
    "y_res = y.copy()\n",
    "\n",
    "# ---------- 3) Split + scaling data to have training and testing splits ----------\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X_res, y_res, test_size=0.25, random_state=42)\n",
    "y_train = y_train.astype(int); y_test = y_test.astype(int)\n",
    "\n",
    "\n",
    "# Scale the variables\n",
    "scaler = StandardScaler()\n",
    "X_train_s = scaler.fit_transform(X_train)\n",
    "X_test_s  = scaler.transform(X_test)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Explore and visualize the feature distributions. \n",
    "* Sample script to visualize the distribution of Volume is shown for training and testing datasets. Do the same for the other continuous attributes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Helper function\n",
    "def plot_by_group(values, groups, title, ax=None):\n",
    "    \"\"\"Overlaid histograms (with density curves) of `values`, one color per group.\"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "    sns.histplot(x=np.asarray(values), hue=np.asarray(groups), kde=True, stat=\"density\",\n",
    "                 common_norm=False, element=\"step\", ax=ax)\n",
    "    ax.set_title(title)\n",
    "    ax.set_xlabel(\"value\")\n",
    "    return ax"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make a 2 × 2 grid of plots. The top row shows D30% and V50Gy for the ORN groups. The bottom row shows the distribution of volume and age by ORN groups. The first panel is done for you. Feature names are \"V50Gy\", \"Mandible Volume\", and \"Age\".\n",
    "What do you notice about the datasets? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1100x700 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(2, 2, figsize=(11, 7))\n",
    "\n",
    "plot_by_group(X['D30%'], y, \"D30%\", ax=axes[0, 0])\n",
    "\n",
    "# TODO: fill in the other three panels -- \n",
    "#plot_by_group(.., y, \"\", ax=axes[0, 1])\n",
    "#plot_by_group(.., .., \"Volume\", ax=axes[1, 0])\n",
    "...\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compute basic statistics for the datasets \n",
    "* Compute mean, SD, Skewness, and kurtosis for the continuous predictors \n",
    "* Do Welch's t test to detemine if the features are significantly different between ORN and no ORN classes\n",
    "* Hint: The methods to calculate stats for one predictor Volume is shown. Do the same for Age, D30% and V50Gy\n",
    "* Hint: for statistical comparison of unparied sets, you need stats.ttest_ind(x1, x2, equal_var=False) if you use Welch's test\n",
    "* "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ORN Volume:- Mean: 86.269898989899 SD : 14.796171255706012 Skewness : 0.0222050809642464 Kurtosis :  -0.3784797118070715\n",
      "No ORN Volume:- Mean: 84.51299677765844 SD : 17.43398670766624 Skewness : -0.040952389365646 Kurtosis :  0.7456077674022739\n",
      "Test volume | P-value: 0.1438798665820787\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Ellipsis"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "import scipy.stats as stats\n",
    "\n",
    "X_ORN = df[df['ORN'] == 1]\n",
    "X_NoORN = df[df['ORN'] == 0]\n",
    "\n",
    "print(\"ORN Volume:- Mean:\", np.mean(X_ORN['Mandible Volume']), \"SD :\",np.std(X_ORN['Mandible Volume']), \"Skewness :\", stats.skew(X_ORN['Mandible Volume']), \"Kurtosis : \", stats.kurtosis(X_ORN['Mandible Volume']))\n",
    "print(\"No ORN Volume:- Mean:\", np.mean(X_NoORN['Mandible Volume']), \"SD :\",np.std(X_NoORN['Mandible Volume']), \"Skewness :\", stats.skew(X_NoORN['Mandible Volume']), \"Kurtosis : \", stats.kurtosis(X_NoORN['Mandible Volume']))\n",
    "\n",
    "t_stat, pval = stats.ttest_ind(X_ORN['Mandible Volume'], X_NoORN['Mandible Volume'], equal_var=False)\n",
    "print(\"Test volume | P-value:\", pval)\n",
    "\n",
    "...\n",
    "...\n",
    "...\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "7LNBdNnU0zyH",
    "outputId": "075191d8-b64a-4408-9a9a-1c1678ccd249"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Visualizing Class Distribution\n",
      "This plot shows how many samples belong to each class (0=Control, 1=ORN).\n",
      "Unique classes found: [0 1]\n",
      "Class counts: {np.int64(0): np.int64(931), np.int64(1): np.int64(198)}\n",
      "Gender\n",
      "Male      938\n",
      "Female    191\n",
      "Name: count, dtype: int64\n",
      "Categories Sex: ['Male', 'Female']\n",
      "Counts: [938, 191]\n",
      "PostRT\n",
      "Definitive    932\n",
      "postop_RT     197\n",
      "Name: count, dtype: int64\n",
      "Categories PostRT: ['Definitive', 'postop_RT']\n",
      "Counts: [932, 197]\n",
      "Site\n",
      "Oropharynx                                      740\n",
      "Oral Cavity                                     196\n",
      "Larynx/Hypopharynx                              163\n",
      "Nasopharynx, Nasal Cavity, Paranasal Sinuses     23\n",
      "Other                                             6\n",
      "Major Salivary Glands                             1\n",
      "Name: count, dtype: int64\n",
      "Categories Tumor: ['Oropharynx', 'Oral Cavity', 'Larynx/Hypopharynx', 'Nasopharynx, Nasal Cavity, Paranasal Sinuses', 'Other', 'Major Salivary Glands']\n",
      "Counts: [740, 196, 163, 23, 6, 1]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x400 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1000x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ======================================================\n",
    "# 3. Visualize your data\n",
    "# ======================================================\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "# ---------- 1) Class distribution ----------\n",
    "print(\"\\n Visualizing Class Distribution\")\n",
    "print(\"This plot shows how many samples belong to each class (0=Control, 1=ORN).\")\n",
    "\n",
    "y = y.astype(int)\n",
    "vals, counts = np.unique(y.values, return_counts=True)\n",
    "\n",
    "# Print out the unique classes and the counts\n",
    "print(f\"Unique classes found: {np.unique(y)}\")\n",
    "print(f\"Class counts: {dict(zip(vals, counts))}\")\n",
    "# Create a fancier bar plot\n",
    "fig, axes = plt.subplots(2, 2, figsize=(10, 4))\n",
    "\n",
    "\n",
    "bars = axes[0,0].bar([str(int(v)) for v in vals], counts,\n",
    "               color=['#1E88E5', '#FF9800'],  # Blue for Control (0), Orange for ORN (1)\n",
    "               edgecolor=['#1565C0', '#EF6C00'],  # Darker borders for contrast\n",
    "               linewidth=1.5, alpha=0.9)  # Slight transparency for polish\n",
    "\n",
    "# Add value labels on top of bars\n",
    "for bar in bars:\n",
    "    height = bar.get_height()\n",
    "    axes[0,0].text(bar.get_x() + bar.get_width() / 2, height + 0.5, f'{int(height)}',\n",
    "             ha='center', va='bottom', fontsize=12, fontweight='bold')\n",
    "\n",
    "axes[0,0].set_title('Class Distribution (0=Control, 1=ORN)', fontsize=12, pad=15, fontweight='bold')\n",
    "axes[0,0].set_xlabel('Class', fontsize=12)\n",
    "axes[0,0].set_ylabel('Count', fontsize=12)\n",
    "axes[0,0].tick_params(axis='both', labelsize=12)\n",
    "\n",
    "## visualize frequency of categorical variables\n",
    "\n",
    "## Do the visualization for Gender\n",
    "counts = X['Gender'].value_counts()\n",
    "print(counts)\n",
    "category_names = counts.index.tolist()\n",
    "category_counts = counts.values.tolist()\n",
    "\n",
    "print(\"Categories Gender:\", category_names)\n",
    "print(\"Counts:\", category_counts)\n",
    "\n",
    "cat_to_int = {cat: i for i, cat in enumerate(category_names)}\n",
    "vals = [cat_to_int[c] for c in category_names]\n",
    "counts = category_counts\n",
    "\n",
    "bars = axes[1,0].bar([str(int(v)) for v in vals], counts,\n",
    "               color=['#1E88E5', '#FF9800'],  # Blue for Control (0), Orange for ORN (1)\n",
    "               edgecolor=['#1565C0', '#EF6C00'],  # Darker borders for contrast\n",
    "               linewidth=1.5, alpha=0.9)  # Slight transparency for polish\n",
    "\n",
    "# Add value labels on top of bars\n",
    "for bar in bars:\n",
    "    height = bar.get_height()\n",
    "    axes[1,0].text(bar.get_x() + bar.get_width() / 2, height + 0.5, f'{int(height)}',\n",
    "             ha='center', va='bottom', fontsize=12, fontweight='bold')\n",
    "\n",
    "axes[1,0].set_title('Gender Distribution (0=Male, 1=Female)', fontsize=16, pad=15, fontweight='bold')\n",
    "axes[1,0].set_xlabel('Gender', fontsize=12)\n",
    "axes[1,0].set_ylabel('Count', fontsize=12)\n",
    "\n",
    "\n",
    "## Do the visualization for PostRT -- Use similar approach as Gender \n",
    "counts = X['PostRT'].value_counts()\n",
    "...\n",
    "\n",
    "# uncomment after you write the code to extract category names and counts for PostRT\n",
    "#print(\"Categories PostRT:\", category_names)\n",
    "#print(\"Counts:\", category_counts)\n",
    "\n",
    "# Plot the PostRT similar to Gender\n",
    "bars = axes[0,1].bar([str(int(v)) for v in vals], counts,\n",
    "               color=['#1E88E5', '#FF9800'],  # Blue for Control (0), Orange for ORN (1)\n",
    "               edgecolor=['#1565C0', '#EF6C00'],  # Darker borders for contrast\n",
    "               linewidth=1.5, alpha=0.9)  # Slight transparency for polish\n",
    "\n",
    "# Add value labels on top of bars\n",
    "for bar in bars:\n",
    "    height = bar.get_height()\n",
    "    axes[0,1].text(bar.get_x() + bar.get_width() / 2, height + 0.5, f'{int(height)}',\n",
    "             ha='center', va='bottom', fontsize=12, fontweight='bold')\n",
    "\n",
    "axes[0,1].set_title('Post-RT (0=na, 1=Post-RT, 2=DRT)', fontsize=16, pad=15, fontweight='bold')\n",
    "axes[0,1].set_xlabel('PostRT', fontsize=12)\n",
    "axes[0,1].set_ylabel('Count', fontsize=12)\n",
    "\n",
    "## Implement the same for tumor sites\n",
    "counts = X['Site'].value_counts()\n",
    "...\n",
    "...\n",
    "...\n",
    "\n",
    "## As there are more sites, we will use this function to get automatically supplied colors\n",
    "colors = np.random.rand(len(counts), 3)\n",
    "\n",
    "bars = axes[1,1].bar([str(int(v)) for v in vals], counts,\n",
    "               color=colors,\n",
    "               edgecolor=colors,\n",
    "               linewidth=1.5, alpha=0.9)  # Slight transparency for polish\n",
    "\n",
    "# Add value labels on top of bars\n",
    "for bar in bars:\n",
    "    height = bar.get_height()\n",
    "    axes[1,1].text(bar.get_x() + bar.get_width() / 2, height + 0.5, f'{int(height)}',\n",
    "             ha='center', va='bottom', fontsize=12, fontweight='bold')\n",
    "\n",
    "axes[1,1].set_title('Tumor Location', fontsize=16, pad=15, fontweight='bold')\n",
    "axes[1,1].set_xlabel('Location', fontsize=12)\n",
    "axes[1,1].set_ylabel('Count', fontsize=12)\n",
    "\n",
    "\n",
    "plt.tight_layout()  # Adjust layout to prevent clipping\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a basic logistic regression classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "pahZxK0G0N5E",
    "outputId": "7e300b83-faae-422c-9d41-42629db283a4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "### Logistic regression classifier ###\n",
      "Intercept: -4.171\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Predictor</th>\n",
       "      <th>Coefficient (log-odds)</th>\n",
       "      <th>Odds ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Site_Nasopharynx, Nasal Cavity, Paranasal Sinuses</td>\n",
       "      <td>-1.1705</td>\n",
       "      <td>0.3102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Site_Major Salivary Glands</td>\n",
       "      <td>0.6210</td>\n",
       "      <td>1.8609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Chemotherapy_Induction Chemotherapy</td>\n",
       "      <td>-0.5298</td>\n",
       "      <td>0.5887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Chemotherapy_No Chemotherapy</td>\n",
       "      <td>-0.5285</td>\n",
       "      <td>0.5895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Gender_Male</td>\n",
       "      <td>0.4375</td>\n",
       "      <td>1.5489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Extractions</td>\n",
       "      <td>0.4078</td>\n",
       "      <td>1.5035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>PostRT_postop_RT</td>\n",
       "      <td>-0.3329</td>\n",
       "      <td>0.7168</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Site_Oral Cavity</td>\n",
       "      <td>0.2955</td>\n",
       "      <td>1.3437</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Site_Oropharynx</td>\n",
       "      <td>0.2813</td>\n",
       "      <td>1.3248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Chemotherapy_Induction and Concurrent</td>\n",
       "      <td>-0.0659</td>\n",
       "      <td>0.9362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>D30%</td>\n",
       "      <td>0.0578</td>\n",
       "      <td>1.0595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Site_Other</td>\n",
       "      <td>0.0512</td>\n",
       "      <td>1.0525</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Age</td>\n",
       "      <td>-0.0185</td>\n",
       "      <td>0.9817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>V50Gy</td>\n",
       "      <td>0.0078</td>\n",
       "      <td>1.0078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Mandible Volume</td>\n",
       "      <td>-0.0011</td>\n",
       "      <td>0.9989</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            Predictor  Coefficient (log-odds)  \\\n",
       "7   Site_Nasopharynx, Nasal Cavity, Paranasal Sinuses                 -1.1705   \n",
       "6                          Site_Major Salivary Glands                  0.6210   \n",
       "11                Chemotherapy_Induction Chemotherapy                 -0.5298   \n",
       "13                       Chemotherapy_No Chemotherapy                 -0.5285   \n",
       "5                                         Gender_Male                  0.4375   \n",
       "4                                         Extractions                  0.4078   \n",
       "14                                   PostRT_postop_RT                 -0.3329   \n",
       "8                                    Site_Oral Cavity                  0.2955   \n",
       "9                                     Site_Oropharynx                  0.2813   \n",
       "12              Chemotherapy_Induction and Concurrent                 -0.0659   \n",
       "0                                                D30%                  0.0578   \n",
       "10                                         Site_Other                  0.0512   \n",
       "3                                                 Age                 -0.0185   \n",
       "1                                               V50Gy                  0.0078   \n",
       "2                                     Mandible Volume                 -0.0011   \n",
       "\n",
       "    Odds ratio  \n",
       "7       0.3102  \n",
       "6       1.8609  \n",
       "11      0.5887  \n",
       "13      0.5895  \n",
       "5       1.5489  \n",
       "4       1.5035  \n",
       "14      0.7168  \n",
       "8       1.3437  \n",
       "9       1.3248  \n",
       "12      0.9362  \n",
       "0       1.0595  \n",
       "10      1.0525  \n",
       "3       0.9817  \n",
       "1       1.0078  \n",
       "2       0.9989  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0      0.830     0.983     0.900       234\n",
      "           1      0.333     0.041     0.073        49\n",
      "\n",
      "    accuracy                          0.820       283\n",
      "   macro avg      0.582     0.512     0.486       283\n",
      "weighted avg      0.744     0.820     0.757       283\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# ============================================\n",
    "# 3. Step 2.2 — Logistic regression\n",
    "# ============================================\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "## Standard random forest classifier without class imbalance or handling large imbalance in predictors prevalence\n",
    "\n",
    "from sklearn.inspection import permutation_importance\n",
    "features_hot = X_train.columns.tolist()\n",
    "\n",
    "print(\"\\n### Logistic regression classifier ###\")\n",
    "simple = LogisticRegression().fit(X_train, y_train)\n",
    "coefficients = pd.DataFrame({\n",
    "    'Predictor': features_hot,\n",
    "    'Coefficient (log-odds)': simple.coef_[0],        # [0] = the single row for class 1 (ORN)\n",
    "    'Odds ratio': np.exp(simple.coef_[0]),            # change in odds of ORN per 1-unit increase\n",
    "}).sort_values('Coefficient (log-odds)', key=abs, ascending=False)\n",
    "\n",
    "print('Intercept:', round(simple.intercept_[0], 4))\n",
    "display(coefficients.round(4))\n",
    "\n",
    "y_pred = simple.predict(X_test)\n",
    "print(classification_report(y_test, y_pred, digits=3))\n",
    "cm = confusion_matrix(y_test, y_pred)\n",
    "tn, fp, fn, tp = cm.ravel()\n",
    "\n",
    "\n",
    "ConfusionMatrixDisplay.from_predictions(y_test, y_pred, cmap='Blues',\n",
    "                                        display_labels=['No ORN', 'ORN'],\n",
    "                                        text_kw={'fontsize': 16, 'fontweight': 'bold'})\n",
    "plt.title('Logistic regression')\n",
    "plt.show()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kuge9jA4B5N7"
   },
   "source": [
    "* What can you tell from the coefficients\n",
    "* What do the accuracies look like?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python [conda env:base] *",
   "language": "python",
   "name": "conda-base-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.13.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}