Decision Tree.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "924f5e8c-40d2-4297-8b13-fca211a98a20",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dataclasses import dataclass\n",
    "\n",
    "@dataclass\n",
    "class Node:\n",
    "    feature: int = None\n",
    "    value: float = None\n",
    "    left: np.array = None\n",
    "    right: np.array = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c4beb19d-8b01-44af-8455-d1357a855090",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.base import BaseEstimator, ClassifierMixin\n",
    "\n",
    "\n",
    "class DecisionTreeClassifier(BaseEstimator, ClassifierMixin):\n",
    "    def __init__(self):\n",
    "        self.root = Node()\n",
    "\n",
    "    @staticmethod\n",
    "    def _gini(y):\n",
    "        \"\"\"Gini impurity.\"\"\"\n",
    "        counts = np.bincount(y)\n",
    "        p = counts / counts.sum()\n",
    "\n",
    "        return (p*(1-p)).sum()\n",
    "\n",
    "    def _split(self, X, y):\n",
    "        \"\"\"Bruteforce search over all features and splitting points.\"\"\"\n",
    "        best_information_gain = float(\"-inf\")\n",
    "        best_feature = None\n",
    "        best_split = None\n",
    "\n",
    "        for feature in range(X.shape[1]):\n",
    "            split_candidates = np.unique(X[:, feature])\n",
    "            for split in split_candidates:\n",
    "                left_mask = X[:, feature] < split\n",
    "                X_left, y_left = X[left_mask], y[left_mask]\n",
    "                X_right, y_right = X[~left_mask], y[~left_mask]\n",
    "\n",
    "                information_gain = self._gini(y) - (\n",
    "                    len(X_left) / len(X) * self._gini(y_left)\n",
    "                    + len(X_right) / len(X) * self._gini(y_right)\n",
    "                )\n",
    "\n",
    "                if information_gain > best_information_gain:\n",
    "                    best_information_gain = information_gain\n",
    "                    best_feature = feature\n",
    "                    best_split = split\n",
    "\n",
    "        return best_feature, best_split\n",
    "\n",
    "    def _build_tree(self, X, y):\n",
    "        \"\"\"The heavy lifting.\"\"\"\n",
    "        feature, split = self._split(X, y)\n",
    "\n",
    "        left_mask = X[:, feature] < split\n",
    "\n",
    "        X_left, y_left = X[left_mask], y[left_mask]\n",
    "        X_right, y_right = X[~left_mask], y[~left_mask]\n",
    "\n",
    "        if len(X_left) == 0 or len(X_right) == 0:\n",
    "            return Node(value=np.argmax(np.bincount(y)))\n",
    "        else:\n",
    "            return Node(\n",
    "                feature,\n",
    "                split,\n",
    "                self._build_tree(X_left, y_left),\n",
    "                self._build_tree(X_right, y_right),\n",
    "            )\n",
    "\n",
    "    def _find_path(self, x, node):\n",
    "        \"\"\"Given a data point x, walk from the root to the corresponding leaf node.\"\"\"\n",
    "        if node.feature == None:\n",
    "            return node.value\n",
    "        else:\n",
    "            if x[node.feature] < node.value:\n",
    "                return self._find_path(x, node.left)\n",
    "            else:\n",
    "                return self._find_path(x, node.right)\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        self.root = self._build_tree(X, y)\n",
    "        return self\n",
    "\n",
    "    def predict(self, X):\n",
    "        return np.array([self._find_path(x, self.root) for x in X])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "202cc9a3-733b-4f10-a903-34e5b2f5ce2c",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "\n",
    "X = np.random.randn(100, 2)\n",
    "y = ((X[:, 0] > 0) * (X[:, 1] < 0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "73d3bfee-3443-4f76-8500-bbdd1bc21677",
   "metadata": {},
   "outputs": [],
   "source": [
    "dt = DecisionTreeClassifier().fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "0f1eb2f5-9856-45e2-b2b1-1a8a753822bd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt.score(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "4205cc4c-73b8-4d0b-bc48-654ff5122d19",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Node(feature=1, value=-0.14963454032767076, left=Node(feature=0, value=0.04575851730144607, left=Node(feature=None, value=0, left=None, right=None), right=Node(feature=None, value=1, left=None, right=None)), right=Node(feature=None, value=0, left=None, right=None))"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt.root"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "3fad41ef-3c09-41dc-b83a-880add2cc72e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from mlxtend.plotting import plot_decision_regions\n",
    "\n",
    "plt.style.use(\"default\")\n",
    "\n",
    "X = np.random.uniform(0, 1, size=(50, 2))\n",
    "y = np.random.randint(5, size=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "04873f4d-c77e-4dc8-9590-456095f05f75",
   "metadata": {},
   "outputs": [],
   "source": [
    "dt = DecisionTreeClassifier().fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "e3461308-b917-4f95-8fc4-409eb3d08931",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\rober\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\mlxtend\\plotting\\decision_regions.py:300: UserWarning: You passed a edgecolor/edgecolors ('black') for an unfilled marker ('x').  Matplotlib is ignoring the edgecolor in favor of the facecolor.  This behavior may change in the future.\n",
      "  ax.scatter(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(-0.1, 1.1)"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 8))\n",
    "plot_decision_regions(X, y, clf=dt, legend=0, hide_spines=True)\n",
    "plt.axis(\"off\")\n",
    "plt.xlim(-0.1, 1.1)\n",
    "plt.ylim(-0.1, 1.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "857c56b0-8d34-4585-b014-fc6d631b1386",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt.score(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "44ccf485-7fcb-490b-bd9f-1f66b3764775",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}