examples/confusion.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('..')\n",
    "\n",
    "import numpy as np\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from pandas import crosstab\n",
    "from korr import confusion"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test 1 - Small Datasets\n",
    "It really does not matter for tiny dataset what you use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(43)\n",
    "y_true = (np.random.normal(size=(100,2)) > 0).astype(int)\n",
    "y_pred = y_true[:, 0]\n",
    "y_true = y_true[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 30.6 ms, sys: 4.2 ms, total: 34.8 ms\n",
      "Wall time: 33.6 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = crosstab(y_true, y_pred)  # pandas\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2.37 ms, sys: 1.45 ms, total: 3.82 ms\n",
      "Wall time: 2.44 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion_matrix(y_true, y_pred) # sklearn\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 183 µs, sys: 59 µs, total: 242 µs\n",
      "Wall time: 215 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion(y_true, y_pred)  # korr\n",
    "del cm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test 2 - Medium Sized Data\n",
    "pandas `crosstab` starts to outperform sklearn's `confusion_matrix`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(43)\n",
    "y_true = (np.random.normal(size=(50000,2)) > 0).astype(int)\n",
    "y_pred = y_true[:, 0]\n",
    "y_true = y_true[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 33.5 ms, sys: 6.31 ms, total: 39.8 ms\n",
      "Wall time: 51.8 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = crosstab(y_true, y_pred)  # pandas\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 76.3 ms, sys: 4.32 ms, total: 80.6 ms\n",
      "Wall time: 106 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion_matrix(y_true, y_pred) # sklearn\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 4.71 ms, sys: 3.6 ms, total: 8.31 ms\n",
      "Wall time: 4.01 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion(y_true, y_pred)  # korr\n",
    "del cm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test 3 - Millions of Obs\n",
    "* sklearn's `confusion_matrix` does not seem to scale well.\n",
    "* pandas' `crosstab` does well. I gues the difference to `korr.confusion` are just error checking routines."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(43)\n",
    "y_true = (np.random.normal(size=(10000000,2)) > 0).astype(int)\n",
    "y_pred = y_true[:, 0]\n",
    "y_true = y_true[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.38 s, sys: 925 ms, total: 2.3 s\n",
      "Wall time: 3.18 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = crosstab(y_true, y_pred)  # pandas\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 12.8 s, sys: 768 ms, total: 13.6 s\n",
      "Wall time: 14.5 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion_matrix(y_true, y_pred) # sklearn\n",
    "del cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 367 ms, sys: 12.8 ms, total: 380 ms\n",
      "Wall time: 385 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cm = confusion(y_true, y_pred)  # korr\n",
    "del cm"
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}