Machine_learning_to_classify_energetic_materials.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Machine learning to classify energetic materials\n",
    "\n",
    "This notebook is based on the tutorial **Your First Machine Learning Project in Python Step-By-Step** (https://machinelearningmastery.com/machine-learning-in-python-step-by-step/), with data from the paper **Applying machine learning techniques to predict the properties of energetic materials** (https://www.nature.com/articles/s41598-018-27344-x).\n",
    "\n",
    "**If you find any mistakes, please let me know**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Introduction\n",
    "\n",
    "A machine learning project may not be linear, but it has a number of well known steps:\n",
    "\n",
    "- Define Problem.\n",
    "- Prepare Data.\n",
    "- Evaluate Algorithms.\n",
    "- Improve Results.\n",
    "- Present Results.\n",
    "\n",
    "The best way to really come to terms with a new platform or tool is to work through a machine learning project end-to-end and cover the key steps. Namely, from loading data, summarizing data, evaluating algorithms and making some predictions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The goal of this notebook is to present a step-by-step, simple project, that uses machine learning in chemistry. We will:\n",
    "\n",
    "- Import a dataset with chemical structures and a response variable of interest.\n",
    "- Calculate chemical descriptors from the structures, using **RDKit**.\n",
    "- Visualize and summarize information in our dataset, using **Pandas** and **Matplotlib**.\n",
    "- Generate predictive classification models, using **Scikit-Learn**.\n",
    "- Estimate the accuracy of our best model using a validation set.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note**: this tutorial originally uses the iris dataset. I adapted it to use a more chemical relevant dataset from the paper \"Applying machine learning techniques to predict the properties of energetic materials\".\n",
    "\n",
    "Link to download the dataset: https://github.com/delton137/Machine-Learning-Energetic-Molecules-Notebooks/blob/master/datasets/Huang_Massa_data_with_all_SMILES.xlsx"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Packages\n",
    "If you need help installing Python or packages, please refer to the original tutorial.\n",
    "\n",
    "In my conda environment, I have already installed the required packages. Several of these come pre-installed on miniconda, but in case you need to install any of them, run this on the environment you are working:\n",
    "\n",
    "- Matplotlib: conda install -c anaconda matplotlib\n",
    "- Numpy: conda install -c anaconda numpy\n",
    "- Pandas: conda install -c anaconda pandas\n",
    "- Scikit-learn: conda install -c anaconda scikit-learn\n",
    "- Scipy: conda install -c anaconda scipy\n",
    "- XLRD (to read .xlsx files): conda install -c anaconda xlrd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python: 3.7.7 (default, May  6 2020, 11:45:54) [MSC v.1916 64 bit (AMD64)]\n",
      "matplotlib: 3.2.1\n",
      "numpy: 1.18.5\n",
      "pandas: 1.0.5\n",
      "rdkit: 2020.03.2\n",
      "sklearn: 0.23.2\n"
     ]
    }
   ],
   "source": [
    "# Python version\n",
    "import sys\n",
    "print('Python: {}'.format(sys.version))\n",
    "# matplotlib\n",
    "import matplotlib\n",
    "print('matplotlib: {}'.format(matplotlib.__version__))\n",
    "# numpy\n",
    "import numpy as np\n",
    "print('numpy: {}'.format(np.__version__))\n",
    "# pandas\n",
    "import pandas as pd\n",
    "print('pandas: {}'.format(pd.__version__))\n",
    "# rdkit\n",
    "import rdkit\n",
    "print('rdkit: {}'.format(rdkit.__version__))\n",
    "# scikit-learn (is imported as sklearn)\n",
    "import sklearn\n",
    "print('sklearn: {}'.format(sklearn.__version__))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Importing other modules and functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from pandas.plotting import scatter_matrix\n",
    "from rdkit import Chem\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing the dataset and a brief overview"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "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>num</th>\n",
       "      <th>Molecular Name</th>\n",
       "      <th>Formula</th>\n",
       "      <th>Density (g/cm3)</th>\n",
       "      <th>Delta Hf solid (kj/mol)</th>\n",
       "      <th>Explosive energy (kj/cc)</th>\n",
       "      <th>Shock velocity (km/s)</th>\n",
       "      <th>Particle velocity (km/s)</th>\n",
       "      <th>Speed of sound (km/s)</th>\n",
       "      <th>Pressure (Gpa)</th>\n",
       "      <th>T(K)</th>\n",
       "      <th>TNT Equiv (per cc)</th>\n",
       "      <th>group</th>\n",
       "      <th>SMILES</th>\n",
       "      <th>NOTE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>CUGCOW</td>\n",
       "      <td>C8N7O14H1</td>\n",
       "      <td>2.024</td>\n",
       "      <td>641.16</td>\n",
       "      <td>6.419</td>\n",
       "      <td>9.569</td>\n",
       "      <td>2.519</td>\n",
       "      <td>7.050</td>\n",
       "      <td>48.777</td>\n",
       "      <td>5440</td>\n",
       "      <td>2.233</td>\n",
       "      <td>Cubane</td>\n",
       "      <td>O=N(=O)C12C3C4(N(=O)=O)C1(N(=O)=O)C1(N(=O)=O)C...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>PUBMUU02</td>\n",
       "      <td>C6N12O12H6</td>\n",
       "      <td>2.044</td>\n",
       "      <td>430.83</td>\n",
       "      <td>5.623</td>\n",
       "      <td>9.972</td>\n",
       "      <td>2.346</td>\n",
       "      <td>7.626</td>\n",
       "      <td>47.807</td>\n",
       "      <td>3886</td>\n",
       "      <td>1.828</td>\n",
       "      <td>CL20</td>\n",
       "      <td>O=N(=O)N1C2C3N(N(=O)=O)C4C1N(N(=O)=O)C(C(N2N(=...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>KURT08MO</td>\n",
       "      <td>C4N8O10H4F4</td>\n",
       "      <td>2.028</td>\n",
       "      <td>-151.80</td>\n",
       "      <td>5.290</td>\n",
       "      <td>8.365</td>\n",
       "      <td>2.286</td>\n",
       "      <td>6.079</td>\n",
       "      <td>38.780</td>\n",
       "      <td>4342</td>\n",
       "      <td>1.842</td>\n",
       "      <td>Linear</td>\n",
       "      <td>[H]C([H])(N(N(=O)=O)C([H])([H])C(N(F)F)(N(=O)=...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>ELUDEV</td>\n",
       "      <td>C4N10O6H2</td>\n",
       "      <td>1.996</td>\n",
       "      <td>657.13</td>\n",
       "      <td>5.228</td>\n",
       "      <td>9.729</td>\n",
       "      <td>2.289</td>\n",
       "      <td>7.441</td>\n",
       "      <td>44.490</td>\n",
       "      <td>3950</td>\n",
       "      <td>1.698</td>\n",
       "      <td>Pyrazole</td>\n",
       "      <td>O=N(=O)NC1=NON=C1N=NC1=NON=C1NN(=O)=O</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>UGUGUY</td>\n",
       "      <td>C8N6O12H2</td>\n",
       "      <td>1.931</td>\n",
       "      <td>593.87</td>\n",
       "      <td>5.188</td>\n",
       "      <td>9.061</td>\n",
       "      <td>2.318</td>\n",
       "      <td>6.743</td>\n",
       "      <td>40.555</td>\n",
       "      <td>5184</td>\n",
       "      <td>1.920</td>\n",
       "      <td>Cubane</td>\n",
       "      <td>O=N(=O)C12C3C4(C1C1(N(=O)=O)C2(N(=O)=O)C3(N(=O...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   num Molecular Name      Formula  Density (g/cm3)  Delta Hf solid (kj/mol)  \\\n",
       "0    1         CUGCOW    C8N7O14H1            2.024                   641.16   \n",
       "1    2       PUBMUU02   C6N12O12H6            2.044                   430.83   \n",
       "2    3       KURT08MO  C4N8O10H4F4            2.028                  -151.80   \n",
       "3    4         ELUDEV    C4N10O6H2            1.996                   657.13   \n",
       "4    5         UGUGUY    C8N6O12H2            1.931                   593.87   \n",
       "\n",
       "   Explosive energy (kj/cc)  Shock velocity (km/s)  Particle velocity (km/s)  \\\n",
       "0                     6.419                  9.569                     2.519   \n",
       "1                     5.623                  9.972                     2.346   \n",
       "2                     5.290                  8.365                     2.286   \n",
       "3                     5.228                  9.729                     2.289   \n",
       "4                     5.188                  9.061                     2.318   \n",
       "\n",
       "   Speed of sound (km/s)  Pressure (Gpa)  T(K)  TNT Equiv (per cc)      group  \\\n",
       "0                  7.050          48.777  5440               2.233    Cubane    \n",
       "1                  7.626          47.807  3886               1.828      CL20    \n",
       "2                  6.079          38.780  4342               1.842    Linear    \n",
       "3                  7.441          44.490  3950               1.698  Pyrazole    \n",
       "4                  6.743          40.555  5184               1.920    Cubane    \n",
       "\n",
       "                                              SMILES NOTE  \n",
       "0  O=N(=O)C12C3C4(N(=O)=O)C1(N(=O)=O)C1(N(=O)=O)C...  NaN  \n",
       "1  O=N(=O)N1C2C3N(N(=O)=O)C4C1N(N(=O)=O)C(C(N2N(=...  NaN  \n",
       "2  [H]C([H])(N(N(=O)=O)C([H])([H])C(N(F)F)(N(=O)=...  NaN  \n",
       "3              O=N(=O)NC1=NON=C1N=NC1=NON=C1NN(=O)=O  NaN  \n",
       "4  O=N(=O)C12C3C4(C1C1(N(=O)=O)C2(N(=O)=O)C3(N(=O...  NaN  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Importing the dataset\n",
    "df = pd.read_excel(r\"./Huang_Massa_data_with_all_SMILES.xlsx\", header=0)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(110, 15)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# See how many rows and columns there are on the dataset\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['num', 'Molecular Name', 'Formula', 'Density (g/cm3)',\n",
       "       'Delta Hf solid (kj/mol)', 'Explosive energy (kj/cc)',\n",
       "       'Shock velocity (km/s)', 'Particle velocity (km/s)',\n",
       "       'Speed of sound (km/s)', 'Pressure (Gpa)', 'T(K)', 'TNT Equiv (per cc)',\n",
       "       'group', 'SMILES', 'NOTE'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Column labels\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some compounds have comments under the 'NOTE' column and will be removed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(104, 15)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Keep only the rows that have 'NaN' under the NOTE column\n",
    "df = df[df['NOTE'].isna()]\n",
    "df.shape  # 6 compounds removed\n",
    "# If you want to see the removed compounds, use: removed = df[~df['NOTE'].isna()]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at the molecular structure of some compounds in our dataset.\n",
    "\n",
    "Using the column 'group', we can see some representative compounds from each group."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Butterfly ',\n",
       " 'CL20 ',\n",
       " 'Cubane ',\n",
       " 'HMX ',\n",
       " 'Ketone ',\n",
       " 'Linear ',\n",
       " 'Pyrazole ',\n",
       " 'RDX ',\n",
       " 'Ring ',\n",
       " 'TNT ']"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# There are 10 groups of explosives\n",
    "sorted(list(df['group'].unique()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.PngImagePlugin.PngImageFile image mode=RGB size=1000x500 at 0x23D5719A608>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Selecting the first occurence of each group\n",
    "cpd_indexes = []\n",
    "for group in sorted(list(df['group'].unique())):\n",
    "    for index, row in df.iterrows():\n",
    "        if row[\"group\"] == group:\n",
    "            cpd_indexes.append(index)\n",
    "            break\n",
    "\n",
    "ms = []\n",
    "for x in cpd_indexes:\n",
    "    smi = df['SMILES'].iloc[x]\n",
    "    ms.append(Chem.MolFromSmiles(smi))\n",
    "\n",
    "img = Chem.Draw.MolsToGridImage(ms, molsPerRow=5, subImgSize=(200, 250), legends=sorted(list(df['group'].unique())))\n",
    "img"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preparing our dataset\n",
    "\n",
    "To make this problem similar to the one in the tutorial, we will need:\n",
    "- A categorical target\n",
    "- Numerical attributes\n",
    "\n",
    "\n",
    "Therefore, in this notebook, we will try to predict if a molecule is a \"strong\" explosive (details below) using molecular descriptors."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Categorical target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visually determine a cutoff for weak/strong explosives\n",
    "target = df['Explosive energy (kj/cc)'].sort_values()\n",
    "plt.scatter(range(len(target)), target)\n",
    "plt.ylabel('Explosive energy (kJ/cc)')\n",
    "plt.plot([0, 100], [4, 4], c=\"red\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use Explosive energy = 4 kJ/cc as a cutoff to classify our compounds as weak or strong explosives. This is arbitrary and I claim no knowledge about relative explosive strength. I just chose this value because it looks like the graph has an inflection point near 4, and about half the compounds will be on each category)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Numerical attributes\n",
    "\n",
    "According to the paper linked above, featurization (selecting the appropriate descriptors) is the most critical part of a ML project with few data points. \n",
    "\n",
    "We will use some of the molecular descriptors suggested on the paper:\n",
    "\n",
    "- Oxygen balance = (100/n atoms) \\* (nO - 2\\*nC - nH/2)\n",
    "- Nitrogen/Carbon ratio = (nN/nC)\n",
    "- Raw counts of C, N, H, F\n",
    "- Presence of certain chemical groups\n",
    "\n",
    "The authors used mmltoolkit (https://github.com/delton137/mmltoolkit) to calculate the descriptors. \n",
    "\n",
    "We will use SMARTS searches instead. SMARTS are used to describe chemical patterns. To learn more:\n",
    "https://www.daylight.com/dayhtml_tutorials/languages/smarts/\n",
    "\n",
    "Several SMARTS can be found at:\n",
    "https://www.daylight.com/dayhtml_tutorials/languages/smarts/smarts_examples.html\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to count the number of matches based on given SMARTS\n",
    "def count_chemical_group(mol, smarts):\n",
    "    return len(mol.GetSubstructMatches(Chem.MolFromSmarts(smarts)))\n",
    "\n",
    "\n",
    "# Oxygen balance 100 as defined on the paper\n",
    "def oxygen_balance_100(mol):\n",
    "    mol = Chem.AddHs(mol)\n",
    "    ob = (100/mol.GetNumAtoms()) * (count_chemical_group(mol, '[#8]') \n",
    "                                    - 2*(count_chemical_group(mol, '[#6]')) \n",
    "                                    - 0.5*(count_chemical_group(mol, '[#1]')))\n",
    "    return float(\"{:.2f}\".format(ob))\n",
    "\n",
    "\n",
    "# Calculate properties for each molecule\n",
    "x = []\n",
    "\n",
    "for m in df[\"SMILES\"]:\n",
    "    mol = Chem.MolFromSmiles(m.strip())  # strip() in case there are empty spaces before/after the string\n",
    "    mol = Chem.AddHs(mol)\n",
    "    ob_100 = oxygen_balance_100(mol)\n",
    "    c = count_chemical_group(mol, '[#6]')\n",
    "    n = count_chemical_group(mol, '[#7]')\n",
    "    h = count_chemical_group(mol, '[#1]')\n",
    "    f = count_chemical_group(mol, '[#9]')\n",
    "    nc_ratio = float(\"{:.2f}\".format(n/c))\n",
    "    cnitro = count_chemical_group(mol, '[$([#6]-[NX3](=O)=O),$([#6]-[NX3+](=O)[O-])]')  # C-NO2\n",
    "    nnitro = count_chemical_group(mol, '[$([#7]-[NX3](=O)=O),$([#7]-[NX3+](=O)[O-])]')  # N-NO2\n",
    "    onitro = count_chemical_group(mol, '[$([#8]-[NX3](=O)=O),$([#8]-[NX3+](=O)[O-])]')  # O-NO2\n",
    "    nitrite = count_chemical_group(mol, '[OX2]-[NX2]=[OX1]')  # O-N=O\n",
    "    cnf = count_chemical_group(mol, '[$([CX3]=[NX2]-[F]),$([CX3]:[NX2]-[F])]')  # C=N-F\n",
    "    coh = count_chemical_group(mol, '[#6][OX2H]')  # C-O-H, includes phenol\n",
    "    n_oxide_o = count_chemical_group(mol, '[$([#7+][OX2][#6]);!$([#7](~[O])~[O]);!$([#7]=[#7])]')  # N-O-C\n",
    "        # Won't hit azoxy, nitro, nitroso, or nitrate.\n",
    "    fulminate = count_chemical_group(mol, '[#6]=[NX2]-[OX1]')  # C=N-O\n",
    "    azo = count_chemical_group(mol, '[#6]-[NX2]=[NX2]')  # C-N=N\n",
    "    amine = count_chemical_group(mol, '[NX3;H2;!$(NC=[!#6]);!$(NC#[!#6])][#6]')  # C-NH2\n",
    "    n_oxide_n = count_chemical_group(mol, '[C][N+]([O-])([C])[C]') \n",
    "        # Paper defines this group as C-N(-O)-C, I believe it is '[C][N+]([O-])([C])[C]'\n",
    "    cf = count_chemical_group(mol, '[#6]-[F]')  # C-F\n",
    "    nitroso = count_chemical_group(mol, '[NX2]=[OX1]')  # N=O\n",
    "    carbonyl = count_chemical_group(mol, '[CX3]=[OX1]')  # C=O\n",
    "    \n",
    "    descriptors = [ob_100, c, n, h, f, nc_ratio, cnitro, nnitro, onitro, nitrite, cnf, coh, \n",
    "                   n_oxide_o, fulminate, azo, amine, n_oxide_n, cf, nitroso, carbonyl]\n",
    "    x.append(descriptors)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scaling\n",
    "Before we can proceed, we have to be sure that the ranges of values for each attribute can be compared. For example, we can't have one column with values ranging from 0 to 100 and another with values ranging from 0 to 1.\n",
    "\n",
    "Therefore, we need to perform scaling or standardization. We will use sklearn.StandardScaler, which scales data to zero mean and unit variance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "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>oxygen_balance_100</th>\n",
       "      <th>c</th>\n",
       "      <th>n</th>\n",
       "      <th>h</th>\n",
       "      <th>f</th>\n",
       "      <th>nc_ratio</th>\n",
       "      <th>cnitro</th>\n",
       "      <th>nnitro</th>\n",
       "      <th>onitro</th>\n",
       "      <th>nitrite</th>\n",
       "      <th>cnf</th>\n",
       "      <th>coh</th>\n",
       "      <th>n_oxide_o</th>\n",
       "      <th>fulminate</th>\n",
       "      <th>azo</th>\n",
       "      <th>amine</th>\n",
       "      <th>n_oxide_n</th>\n",
       "      <th>cf</th>\n",
       "      <th>nitroso</th>\n",
       "      <th>carbonyl</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-8.33</td>\n",
       "      <td>8.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.88</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-8.33</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.00</td>\n",
       "      <td>4.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.00</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-13.64</td>\n",
       "      <td>4.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-17.86</td>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.75</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>-77.14</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.17</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>-42.31</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>-50.00</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>-50.00</td>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>-33.33</td>\n",
       "      <td>2.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>104 rows × 20 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     oxygen_balance_100     c     n     h    f  nc_ratio  cnitro  nnitro  \\\n",
       "0                 -8.33   8.0   7.0   1.0  0.0      0.88     7.0     0.0   \n",
       "1                 -8.33   6.0  12.0   6.0  0.0      2.00     0.0     6.0   \n",
       "2                  0.00   4.0   8.0   4.0  4.0      2.00     2.0     1.0   \n",
       "3                -13.64   4.0  10.0   2.0  0.0      2.50     0.0     2.0   \n",
       "4                -17.86   8.0   6.0   2.0  0.0      0.75     6.0     0.0   \n",
       "..                  ...   ...   ...   ...  ...       ...     ...     ...   \n",
       "99               -77.14  12.0   2.0  16.0  0.0      0.17     2.0     0.0   \n",
       "100              -42.31   2.0   5.0   5.0  0.0      2.50     0.0     0.0   \n",
       "101              -50.00   4.0   6.0   2.0  0.0      1.50     0.0     0.0   \n",
       "102              -50.00   4.0   6.0   2.0  0.0      1.50     0.0     0.0   \n",
       "103              -33.33   2.0  10.0   0.0  0.0      5.00     0.0     0.0   \n",
       "\n",
       "     onitro  nitrite  cnf  coh  n_oxide_o  fulminate  azo  amine  n_oxide_n  \\\n",
       "0       0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "1       0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "2       0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "3       0.0      0.0  0.0  0.0        0.0        0.0  2.0    0.0        0.0   \n",
       "4       0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "..      ...      ...  ...  ...        ...        ...  ...    ...        ...   \n",
       "99      0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "100     0.0      0.0  0.0  0.0        0.0        0.0  2.0    0.0        0.0   \n",
       "101     0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "102     0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "103     0.0      0.0  0.0  0.0        0.0        0.0  0.0    0.0        0.0   \n",
       "\n",
       "      cf  nitroso  carbonyl  \n",
       "0    0.0      0.0       0.0  \n",
       "1    0.0      0.0       0.0  \n",
       "2    0.0      0.0       0.0  \n",
       "3    0.0      0.0       0.0  \n",
       "4    0.0      0.0       0.0  \n",
       "..   ...      ...       ...  \n",
       "99   0.0      0.0       0.0  \n",
       "100  0.0      0.0       0.0  \n",
       "101  0.0      0.0       0.0  \n",
       "102  0.0      0.0       0.0  \n",
       "103  0.0      0.0       0.0  \n",
       "\n",
       "[104 rows x 20 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create dataframe with attributes before scaling\n",
    "x = pd.DataFrame(np.array(x), columns=[\"oxygen_balance_100\", \"c\", \"n\", \"h\", \"f\", \"nc_ratio\", \n",
    "                                       \"cnitro\", \"nnitro\", \"onitro\", \"nitrite\", \"cnf\", \n",
    "                                       \"coh\", \"n_oxide_o\", \"fulminate\", \"azo\", \"amine\", \n",
    "                                       \"n_oxide_n\", \"cf\", \"nitroso\", \"carbonyl\"])\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scaling\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "X = pd.DataFrame(StandardScaler().fit_transform(x), columns=[\"oxygen_balance_100\", \"c\", \"n\", \"h\", \"f\", \"nc_ratio\", \n",
    "                                       \"cnitro\", \"nnitro\", \"onitro\", \"nitrite\", \"cnf\", \n",
    "                                       \"coh\", \"n_oxide_o\", \"fulminate\", \"azo\", \"amine\", \n",
    "                                       \"n_oxide_n\", \"cf\", \"nitroso\", \"carbonyl\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Categorization of target\n",
    "Finally, we need to categorize our target column. Every value with Explosive energy > 4 kJ/cc will be assigned the value 1 and categorized as a strong explosive, while values below 4 kJ/cc will be assigned 0.\n",
    "\n",
    "To achieve this, we will use Pandas boolean logic to assign True/False based on the condition Explosive energy > 4, then convert these booleans to integers simply changing the type of the array (Python converts False to 0 and True to 1)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = df[\"Explosive energy (kj/cc)\"] > 4\n",
    "y = y.to_frame()\n",
    "y.rename(columns={\"Explosive energy (kj/cc)\":\"Strong_explosive\"}, inplace = True)\n",
    "Y = y.astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Looking at our dataset\n",
    "Now that our dataset is similar to the one used in the tutorial, we will proceed as suggested. Let's take a look at:\n",
    "- First rows of each dataset.\n",
    "- Dimensions of the dataset.\n",
    "- Peek at the data itself.\n",
    "- Statistical summary of all attributes.\n",
    "- Breakdown of the data by the class variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "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>oxygen_balance_100</th>\n",
       "      <th>c</th>\n",
       "      <th>n</th>\n",
       "      <th>h</th>\n",
       "      <th>f</th>\n",
       "      <th>nc_ratio</th>\n",
       "      <th>cnitro</th>\n",
       "      <th>nnitro</th>\n",
       "      <th>onitro</th>\n",
       "      <th>nitrite</th>\n",
       "      <th>cnf</th>\n",
       "      <th>coh</th>\n",
       "      <th>n_oxide_o</th>\n",
       "      <th>fulminate</th>\n",
       "      <th>azo</th>\n",
       "      <th>amine</th>\n",
       "      <th>n_oxide_n</th>\n",
       "      <th>cf</th>\n",
       "      <th>nitroso</th>\n",
       "      <th>carbonyl</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.561462</td>\n",
       "      <td>1.054682</td>\n",
       "      <td>0.220151</td>\n",
       "      <td>-1.338662</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>-0.625440</td>\n",
       "      <td>3.031064</td>\n",
       "      <td>-0.762365</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.561462</td>\n",
       "      <td>0.245184</td>\n",
       "      <td>2.008878</td>\n",
       "      <td>0.405960</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>0.521102</td>\n",
       "      <td>-0.937667</td>\n",
       "      <td>3.169168</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.921407</td>\n",
       "      <td>-0.564313</td>\n",
       "      <td>0.577897</td>\n",
       "      <td>-0.291889</td>\n",
       "      <td>1.938600</td>\n",
       "      <td>0.521102</td>\n",
       "      <td>0.196256</td>\n",
       "      <td>-0.107109</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.332012</td>\n",
       "      <td>-0.564313</td>\n",
       "      <td>1.293387</td>\n",
       "      <td>-0.989738</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>1.032950</td>\n",
       "      <td>-0.937667</td>\n",
       "      <td>0.548146</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.149663</td>\n",
       "      <td>1.054682</td>\n",
       "      <td>-0.137594</td>\n",
       "      <td>-0.989738</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>-0.758521</td>\n",
       "      <td>2.464103</td>\n",
       "      <td>-0.762365</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-0.101824</td>\n",
       "      <td>1.054682</td>\n",
       "      <td>0.577897</td>\n",
       "      <td>-0.989738</td>\n",
       "      <td>6.613716</td>\n",
       "      <td>-0.502596</td>\n",
       "      <td>-0.937667</td>\n",
       "      <td>1.858657</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.382279</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.921407</td>\n",
       "      <td>1.054682</td>\n",
       "      <td>0.577897</td>\n",
       "      <td>-1.687587</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>-0.502596</td>\n",
       "      <td>3.598026</td>\n",
       "      <td>-0.762365</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.601215</td>\n",
       "      <td>-0.159564</td>\n",
       "      <td>-0.495340</td>\n",
       "      <td>-0.291889</td>\n",
       "      <td>1.354210</td>\n",
       "      <td>-0.502596</td>\n",
       "      <td>0.763218</td>\n",
       "      <td>-0.762365</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.876630</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.990977</td>\n",
       "      <td>1.864180</td>\n",
       "      <td>2.724369</td>\n",
       "      <td>1.801658</td>\n",
       "      <td>0.769821</td>\n",
       "      <td>-0.093117</td>\n",
       "      <td>1.330179</td>\n",
       "      <td>0.548146</td>\n",
       "      <td>3.176198</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.153780</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.761527</td>\n",
       "      <td>-0.159564</td>\n",
       "      <td>1.293387</td>\n",
       "      <td>-0.989738</td>\n",
       "      <td>-0.398958</td>\n",
       "      <td>0.521102</td>\n",
       "      <td>0.196256</td>\n",
       "      <td>0.548146</td>\n",
       "      <td>-0.194461</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.132712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.228086</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.291920</td>\n",
       "      <td>-0.098533</td>\n",
       "      <td>-0.454118</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   oxygen_balance_100         c         n         h         f  nc_ratio  \\\n",
       "0            0.561462  1.054682  0.220151 -1.338662 -0.398958 -0.625440   \n",
       "1            0.561462  0.245184  2.008878  0.405960 -0.398958  0.521102   \n",
       "2            0.921407 -0.564313  0.577897 -0.291889  1.938600  0.521102   \n",
       "3            0.332012 -0.564313  1.293387 -0.989738 -0.398958  1.032950   \n",
       "4            0.149663  1.054682 -0.137594 -0.989738 -0.398958 -0.758521   \n",
       "5           -0.101824  1.054682  0.577897 -0.989738  6.613716 -0.502596   \n",
       "6            0.921407  1.054682  0.577897 -1.687587 -0.398958 -0.502596   \n",
       "7            0.601215 -0.159564 -0.495340 -0.291889  1.354210 -0.502596   \n",
       "8            0.990977  1.864180  2.724369  1.801658  0.769821 -0.093117   \n",
       "9            0.761527 -0.159564  1.293387 -0.989738 -0.398958  0.521102   \n",
       "\n",
       "     cnitro    nnitro    onitro  nitrite  cnf       coh  n_oxide_o  fulminate  \\\n",
       "0  3.031064 -0.762365 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "1 -0.937667  3.169168 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "2  0.196256 -0.107109 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "3 -0.937667  0.548146 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "4  2.464103 -0.762365 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "5 -0.937667  1.858657 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "6  3.598026 -0.762365 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "7  0.763218 -0.762365 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "8  1.330179  0.548146  3.176198      0.0  0.0 -0.132712        0.0        0.0   \n",
       "9  0.196256  0.548146 -0.194461      0.0  0.0 -0.132712        0.0        0.0   \n",
       "\n",
       "   azo     amine  n_oxide_n        cf   nitroso  carbonyl  \n",
       "0 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "1 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "2 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "3  5.0 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "4 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "5 -0.2 -0.228086        0.0  8.382279 -0.098533 -0.454118  \n",
       "6 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  \n",
       "7 -0.2 -0.228086        0.0  1.876630 -0.098533 -0.454118  \n",
       "8 -0.2 -0.228086        0.0  1.153780 -0.098533 -0.454118  \n",
       "9 -0.2 -0.228086        0.0 -0.291920 -0.098533 -0.454118  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# First rows of our scaled descriptors dataset (X)\n",
    "X.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "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>Strong_explosive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Strong_explosive\n",
       "0                  1\n",
       "1                  1\n",
       "2                  1\n",
       "3                  1\n",
       "4                  1\n",
       "54                 0\n",
       "55                 0\n",
       "56                 0\n",
       "57                 0\n",
       "58                 0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Some rows of our responses dataset (Y)\n",
    "Y.iloc[[0, 1, 2, 3, 4, 54, 55, 56, 57, 58]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(104, 20)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dimensions of X dataset\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "       oxygen_balance_100             c             n             h  \\\n",
      "count        1.040000e+02  1.040000e+02  1.040000e+02  1.040000e+02   \n",
      "mean         9.180690e-17  8.540177e-18 -1.579933e-16 -5.124106e-17   \n",
      "std          1.004843e+00  1.004843e+00  1.004843e+00  1.004843e+00   \n",
      "min         -4.479938e+00 -1.778560e+00 -2.284067e+00 -1.687587e+00   \n",
      "25%         -3.390508e-01 -5.643134e-01 -8.530854e-01 -6.408133e-01   \n",
      "50%          1.753734e-01 -1.595645e-01 -1.375944e-01 -2.918888e-01   \n",
      "75%          6.154750e-01  6.499334e-01  5.778965e-01  4.931914e-01   \n",
      "max          2.649838e+00  2.673678e+00  3.439860e+00  3.895206e+00   \n",
      "\n",
      "                  f      nc_ratio        cnitro        nnitro        onitro  \\\n",
      "count  1.040000e+02  1.040000e+02  1.040000e+02  1.040000e+02  1.040000e+02   \n",
      "mean  -4.270089e-18  2.562053e-17  7.686159e-17 -5.978124e-17  3.416071e-17   \n",
      "std    1.004843e+00  1.004843e+00  1.004843e+00  1.004843e+00  1.004843e+00   \n",
      "min   -3.989583e-01 -1.526294e+00 -9.376674e-01 -7.623645e-01 -1.944611e-01   \n",
      "25%   -3.989583e-01 -6.535916e-01 -9.376674e-01 -7.623645e-01 -1.944611e-01   \n",
      "50%   -3.989583e-01 -9.311712e-02  1.962560e-01 -7.623645e-01 -1.944611e-01   \n",
      "75%   -3.989583e-01  3.905801e-01  1.962560e-01  5.481464e-01 -1.944611e-01   \n",
      "max    6.613716e+00  4.615893e+00  3.598026e+00  3.169168e+00  6.546858e+00   \n",
      "\n",
      "       nitrite    cnf           coh  n_oxide_o  fulminate           azo  \\\n",
      "count    104.0  104.0  1.040000e+02      104.0      104.0  1.040000e+02   \n",
      "mean       0.0    0.0 -5.978124e-17        0.0        0.0 -8.540177e-18   \n",
      "std        0.0    0.0  1.004843e+00        0.0        0.0  1.004843e+00   \n",
      "min        0.0    0.0 -1.327122e-01        0.0        0.0 -2.000000e-01   \n",
      "25%        0.0    0.0 -1.327122e-01        0.0        0.0 -2.000000e-01   \n",
      "50%        0.0    0.0 -1.327122e-01        0.0        0.0 -2.000000e-01   \n",
      "75%        0.0    0.0 -1.327122e-01        0.0        0.0 -2.000000e-01   \n",
      "max        0.0    0.0  9.068666e+00        0.0        0.0  5.000000e+00   \n",
      "\n",
      "              amine  n_oxide_n            cf       nitroso      carbonyl  \n",
      "count  1.040000e+02      104.0  1.040000e+02  1.040000e+02  1.040000e+02  \n",
      "mean  -2.135044e-17        0.0  2.562053e-17  1.708035e-17  6.832142e-17  \n",
      "std    1.004843e+00        0.0  1.004843e+00  1.004843e+00  1.004843e+00  \n",
      "min   -2.280858e-01        0.0 -2.919202e-01 -9.853293e-02 -4.541184e-01  \n",
      "25%   -2.280858e-01        0.0 -2.919202e-01 -9.853293e-02 -4.541184e-01  \n",
      "50%   -2.280858e-01        0.0 -2.919202e-01 -9.853293e-02 -4.541184e-01  \n",
      "75%   -2.280858e-01        0.0 -2.919202e-01 -9.853293e-02 -4.541184e-01  \n",
      "max    7.678888e+00        0.0  8.382279e+00  1.014889e+01  3.044275e+00  \n"
     ]
    }
   ],
   "source": [
    "# Statistical summary of X\n",
    "print(X.describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['nitrite', 'cnf', 'n_oxide_o', 'fulminate', 'n_oxide_n']\n"
     ]
    }
   ],
   "source": [
    "# Let's see which columns have all repeated values \n",
    "zero_var_columns = [col for col in X.columns if len(X[col].unique()) == 1]\n",
    "print(zero_var_columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# These can be removed, as they won't contribute to the models\n",
    "X.drop(labels=zero_var_columns, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    54\n",
       "0    50\n",
       "Name: Strong_explosive, dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Class Distribution: how many compounds belong to each class\n",
    "Y['Strong_explosive'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 50 \"weak\" and 54 \"strong\" explosives."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Visualization\n",
    "\n",
    "### Univariate Plots\n",
    "\n",
    "We start with some univariate plots, that is, plots of each individual variable.\n",
    "\n",
    "Given that the input variables are numeric, we can create box and whisker plots of each.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X.boxplot(vert=False, figsize=(20,10), fontsize=22)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note**: remember that we scaled our data, so the negative medians don't mean negative counts of atoms/groups."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Histogram x bar plot\n",
    "\n",
    "The original tutorial shows a histogram for each variable. However, since we are using counts instead of continuous variables, I don't think a histogram would be very informative.\n",
    "\n",
    "Instead, let's make a bar plot to see how the original dataset is divided by groups of explosives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Matplotlib plotting parameters: figure and font size\n",
    "plt.rcParams[\"figure.figsize\"] = (18,10)\n",
    "plt.rcParams['font.size'] = 22\n",
    "\n",
    "group_counts = df[\"group\"].value_counts()\n",
    "plt.barh(group_counts.index, group_counts)\n",
    "plt.xticks(np.arange(0, 21, 1))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Due to this distribution, we can expect that our models will be better classifiers of pyrazole-type compounds than of HMX, CL20 and RDX-type. This could be corrected by adding more examples from the less represented classes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multivariate Plots\n",
    "\n",
    "Now we can look at the interactions between the variables.\n",
    "\n",
    "First, let’s look at scatterplots of some pairs of attributes. This can be helpful to spot structured relationships between input variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 36 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# scatter plot matrix\n",
    "scatter_df = X[[\"oxygen_balance_100\", \"c\", \"n\", \"h\", \"f\", \"nc_ratio\"]]\n",
    "scatter_df.columns = [\"OB100\", \"C\", \"N\", \"H\", \"F\", \"N/C ratio\"]\n",
    "scatter_matrix(scatter_df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It does not look like there are significant correlations between the variables displayed.\n",
    "\n",
    "We could determine if there is multicollinearity in our dataset by calculating the correlation (R, R²) for all pairs of variables. I will leave this to another notebook."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluate Some Algorithms\n",
    "\n",
    "Now, the machine learning part begins. We will:\n",
    "\n",
    "- Separate out a validation dataset. The validation set, or external test set, will be used to estimate the model accuracy in real life.\n",
    "- Set-up the test harness to use 10-fold cross validation.\n",
    "- Build multiple models, compare and select the best model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert datasets to arrays, to avoid Warnings\n",
    "X_array = X.values[:, :]\n",
    "Y_array = Y.to_numpy(dtype=int)\n",
    "\n",
    "# Split dataset into training and validation\n",
    "X_train, X_validation, Y_train, Y_validation = train_test_split(X_array, Y_array, test_size=0.20, random_state=1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we will build models using 7 different algorithms:\n",
    "\n",
    "- Logistic Regression (LR)\n",
    "- K-Nearest Neighbors (KNN).\n",
    "- Gaussian Naive Bayes (NB).\n",
    "- Support Vector Machines (SVM).\n",
    "- Linear Discriminant Analysis (LDA)\n",
    "- Classification and Regression Trees (CART).\n",
    "- Random forests (RF).\n",
    "\n",
    "We will use stratified 10-fold cross validation to estimate model accuracy.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR: 0.81 (0.10)\n",
      "KNN: 0.75 (0.06)\n",
      "NB: 0.59 (0.12)\n",
      "SVM: 0.76 (0.08)\n",
      "LDA: 0.77 (0.06)\n",
      "CART: 0.80 (0.13)\n",
      "RF: 0.79 (0.20)\n"
     ]
    }
   ],
   "source": [
    "# Select Algorithms\n",
    "models = []\n",
    "models.append(('LR', LogisticRegression(solver='liblinear')))\n",
    "    # For small datasets, ‘liblinear’ is a good choice\n",
    "models.append(('KNN', KNeighborsClassifier()))\n",
    "models.append(('NB', GaussianNB()))\n",
    "models.append(('SVM', SVC(gamma='auto')))\n",
    "models.append(('LDA', LinearDiscriminantAnalysis()))\n",
    "models.append(('CART', DecisionTreeClassifier()))\n",
    "models.append(('RF', RandomForestClassifier(random_state=1)))\n",
    "\n",
    "# Evaluate each model with stratified 10-fold cross-validation\n",
    "results = []\n",
    "names = []\n",
    "for name, model in models:\n",
    "    kfold = StratifiedKFold(n_splits=10, random_state=1, shuffle=True)\n",
    "    cv_results = cross_val_score(model, X_train, Y_train.ravel(), cv=kfold, scoring='accuracy')\n",
    "    results.append(cv_results)\n",
    "    names.append(name)\n",
    "    print(f'{name}: {cv_results.mean():.2f} ({cv_results.std():.2f})')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Compare Algorithms\n",
    "plt.boxplot(results, labels=names)\n",
    "plt.title('Algorithm Comparison')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation\n",
    "Now, we choose the best model to apply on the validation set and try to estimate its true accuracy.\n",
    "\n",
    "LR, CART and RF scored higher. While CART and RF have a low-accuracy outlier, LR seems more stable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Validation accuracy score: 0.76\n",
      "\n",
      "Confusion matrix:\n",
      "[[11  2]\n",
      " [ 3  5]]\n",
      "\n",
      "Classification report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.85      0.81        13\n",
      "           1       0.71      0.62      0.67         8\n",
      "\n",
      "    accuracy                           0.76        21\n",
      "   macro avg       0.75      0.74      0.74        21\n",
      "weighted avg       0.76      0.76      0.76        21\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Make predictions on validation dataset\n",
    "model = LogisticRegression(solver='liblinear')\n",
    "model.fit(X_train, Y_train.ravel())\n",
    "predictions = model.predict(X_validation)\n",
    "\n",
    "# Evaluate predictions\n",
    "print(f\"Validation accuracy score: {accuracy_score(Y_validation, predictions):.2f}\")\n",
    "print()\n",
    "print(\"Confusion matrix:\")\n",
    "print(confusion_matrix(Y_validation, predictions))\n",
    "print()\n",
    "# Confusion matrix: https://en.wikipedia.org/wiki/Confusion_matrix\n",
    "print(\"Classification report:\")\n",
    "print(classification_report(Y_validation, predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that even with a model as simple as logistic regression, as well as few descriptors, we were able to achieve a reasonable predictor, with 76% accuracy.\n",
    "\n",
    "This means that the relative strength of an explosive can, to some extent, be predicted by the chemical features defined.\n",
    "\n",
    "Let's see if we can learn anything from the misclassified molecules:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.PngImagePlugin.PngImageFile image mode=RGB size=900x700 at 0x23D57E322C8>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Finding all false negatives and false positives\n",
    "fn = []\n",
    "fp = []\n",
    "ind = 0\n",
    "for i in range(len(Y_validation)):\n",
    "    if Y_validation[i] and (Y_validation[i] != predictions[i]):  # False negative\n",
    "        fn.append(X_validation[ind])\n",
    "    elif predictions[i] and (Y_validation[i] != predictions[i]):  # False positive\n",
    "        fp.append(X_validation[ind])\n",
    "    ind += 1\n",
    "\n",
    "# Finding the indexes that correspond to FNs and FPs\n",
    "indexes = []\n",
    "for value in fn+fp:\n",
    "    for index, row in X.iterrows():\n",
    "        if np.array_equal(value, row):\n",
    "            indexes.append(index)\n",
    "\n",
    "# Show structures of misclassified compounds\n",
    "# Indexes are preserved for every transformation, so indexes in X are the same as in df\n",
    "ms = []\n",
    "for i in indexes:\n",
    "    smi = df['SMILES'].loc[i]  \n",
    "    ms.append(Chem.MolFromSmiles(smi))\n",
    "\n",
    "img_legends = len(fn)*['False Negative'] + len(fp)*['False Positive']\n",
    "img = Chem.Draw.MolsToGridImage(ms, molsPerRow=3, subImgSize=(300, 350), legends=img_legends)\n",
    "img"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The structures don't seem to have anything in common. Let's see to which group they belong:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "35      Linear \n",
      "31      Linear \n",
      "46    Pyrazole \n",
      "59         RDX \n",
      "56         RDX \n",
      "Name: group, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print(df.iloc[indexes]['group'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As stated earlier, we would expect our model to perform worse for RDX, since we have few examples of this class in the dataset. However, Linear and Pyrazole were the groups with the most examples, so maybe the characteristics that define a group are not the same that define explosive strength.\n",
    "\n",
    "We can verify this:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "strong_group_counts = dict()\n",
    "weak_group_counts = dict()\n",
    "\n",
    "for index in group_counts.index:  # This is to make sure that our bar plot will have the same order as the one above\n",
    "    strong_group_counts[index] = 0\n",
    "    weak_group_counts[index] = 0\n",
    "\n",
    "for i in range(len(Y)):\n",
    "    group = df.iloc[i][\"group\"]\n",
    "    if Y.iloc[i][\"Strong_explosive\"]:\n",
    "        if group in strong_group_counts.keys():\n",
    "            strong_group_counts[group] += 1\n",
    "        else:\n",
    "            strong_group_counts[group] = 1\n",
    "    else:\n",
    "        if group in weak_group_counts.keys():\n",
    "            weak_group_counts[group] += 1\n",
    "        else:\n",
    "            weak_group_counts[group] = 1 \n",
    "\n",
    "plt.barh(list(strong_group_counts.keys()), list(strong_group_counts.values()))\n",
    "plt.barh(list(weak_group_counts.keys()), list(weak_group_counts.values()), \n",
    "         left=list(strong_group_counts.values()), color=\"#add8e6\")\n",
    "plt.legend(labels=[\"Strong\", \"Weak\"])\n",
    "plt.xticks(np.arange(0, 21, 1))\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apparently, Linear, Butterfly and Ketone explosives are more likely to be \"strong\", while TNT and Ring, more likely to be \"weak\"\n",
    "\n",
    "Our model was built with several \"strong\" Linear explosives, therefore, it is not surprising that two of the false positives were Linear: Linear explosives must have features associated with \"strong\" explosives.\n",
    "\n",
    "As for Pyrazole and RDX explosives, the distribution of \"strong\":\"weak\" in the dataset was close to 50:50, so this analysis doesn't tell us much about why they were misclassified.\n",
    "\n",
    "However, since the dataset is ordered from stronger to weaker explosive, we can see that the RDX and Pyrazole misclassified explosives are near the middle of the dataset (indexes 46, 56 and 59). This means that their explosive strength is close to our arbitrary boundary of 4 kJ/cc, and it is expected that any model will be better at classifying extremes than compounds near a boundary."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Next steps\n",
    "\n",
    "There are several approaches we could take do to make improved models. For example, we could:\n",
    "1. Add more data\n",
    " - This is the main factor that influences machine learning model. If you can, get more (quality) data. The paper cited above used another dataset with more molecules and it is available on the author's GitHub.\n",
    " \n",
    "2. Calculate more descriptors\n",
    " - The number of chemical groups we could define is immense. However, having more descriptors can lead to problems if we don't add more data. Dimensionality reduction methods could be applied if we had, for instance, more descriptors than data points.\n",
    " - Another way to convert molecules to a machine learning-readable format is using fingerprints. RDKit offers functions to generate several types of fingerprints (MACCS, ECFP, Avalon, E-state, to name a few).\n",
    "  - The authors of the paper used more elaborate descriptors, besides fingerprints, such as: sum over bonds, Coulomb matrices and Bag of Bonds, some of which performed better than chemical descriptors.\n",
    "  \n",
    "3. Use a more complex algorithm\n",
    " - This can be effective if you have more data. Examples of algorithms are the different flavors of neural networks. Maybe I will do another notebook with NNs in the future."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The end!\n",
    "Below, some tips about saving an ML model and some notes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Save the model\n",
    "To save the best model to use in the future, we can use the python module pickle."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "# save the model to disk\n",
    "filename = 'final_model.sav'\n",
    "pickle.dump(model, open(filename, 'wb'))\n",
    " \n",
    "# some time later...\n",
    " \n",
    "# load the model from disk\n",
    "loaded_model = pickle.load(open(filename, 'rb'))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Important: Tips for Saving Your Model**\n",
    "\n",
    "This section lists some important considerations when finalizing your machine learning models.\n",
    "\n",
    "**Python Version**. Take note of the python version. You almost certainly require the same major (and maybe minor) version of Python used to serialize the model when you later load it and deserialize it.\n",
    "\n",
    "**Library Versions**. The version of all major libraries used in your machine learning project almost certainly need to be the same when deserializing a saved model. This is not limited to the version of NumPy and the version of scikit-learn.\n",
    "\n",
    "**Manual Serialization**. You might like to manually output the parameters of your learned model so that you can use them directly in scikit-learn or another platform in the future. Often the algorithms used by machine learning algorithms to make predictions are a lot simpler than those used to learn the parameters can may be easy to implement in custom code that you have control over.\n",
    "\n",
    "Take note of the version so that you can re-create the environment if for some reason you cannot reload your model on another machine or another platform at a later time."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Note: for molecules with >1000 atoms, my count method could yield incorrect values\n",
    "# See: https://sourceforge.net/p/rdkit/mailman/rdkit-discuss/thread/CAFdOP2uSPZv6wgNY9ZRumLB%2Bwxxu%2BLpMyTs8sC_kbn0Pr5YxEw%40mail.gmail.com/#msg34524687\n",
    "# Suggestions for datasets with larger molecules:\n",
    "\n",
    "pat = Chem.MolFromSmarts(\"[#6]\")\n",
    "def count3(mol):\n",
    "    return len(mol.GetSubstructMatches(pat, maxMatches=mol.GetNumAtoms()))\n",
    "\n",
    "from rdkit.Chem import rdqueries\n",
    "q = rdqueries.AtomNumEqualsQueryAtom(6)\n",
    "len(mol.GetAtomsMatchingQuery(q))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drawing in RDKit\n",
    "Chem.Draw.MolsToGridImage(subImgSize=(300, 350)) # looks nice for 3 molecules per row"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Class Distribution merging the X and Y datasets\n",
    "merged = pd.concat([X, Y], axis=1)\n",
    "print(merged.groupby('Strong_explosive').size())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}