Untitled5.ipynb

{
 "cells": [
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "import numpy as np",
   "execution_count": 1,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "truths = np.array([0, 0, 1, 1, 0], dtype=np.int)\npredictions = np.ones(shape=(5,2), dtype=np.float)",
   "execution_count": 2,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictions[:, 1] = 0\npredictions[[2,3], 0] = 0\npredictions[[2,3], 1] = 1",
   "execution_count": 3,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictions",
   "execution_count": 4,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 4,
     "data": {
      "text/plain": "array([[1., 0.],\n       [1., 0.],\n       [0., 1.],\n       [0., 1.],\n       [1., 0.]])"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "import proclam",
   "execution_count": 5,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "D = proclam.metrics.LogLoss()",
   "execution_count": 6,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "D.evaluate(predictions, truths)",
   "execution_count": 7,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 7,
     "data": {
      "text/plain": "2.220446049250313e-16"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "D.evaluate(predictions, truths, averaging='per_class')",
   "execution_count": 8,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 8,
     "data": {
      "text/plain": "2.220446049250313e-16"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Since all predictions are correct, we should expect the logloss to be\n\n5 * log(p=1) = 0"
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictions",
   "execution_count": 9,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 9,
     "data": {
      "text/plain": "array([[1.00000000e+00, 2.22044605e-16],\n       [1.00000000e+00, 2.22044605e-16],\n       [2.22044605e-16, 1.00000000e+00],\n       [2.22044605e-16, 1.00000000e+00],\n       [1.00000000e+00, 2.22044605e-16]])"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Assuming this was clipped so that the probabilities don't add up to 1.0 but 1.0 + 2.22e-16, \nThis would be -5 * log(1-2.2e-16)/ 5 = 2.2e-16 . "
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "newpredictions = np.ones(shape=(5,2), dtype=np.float)",
   "execution_count": 10,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "newpredictions[:, 1] = 0",
   "execution_count": 11,
   "outputs": []
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "D.evaluate(newpredictions, truths, averaging='per_class')",
   "execution_count": 21,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 21,
     "data": {
      "text/plain": "18.021826694558577"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "Now we have two incorrect ones, so I would expect \n(- 3* log(1)/3. - 2*log(2.2e-16)/2. -5* 2.2e-16) / 5"
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "2 * 16 *np.log(2.2) /2.",
   "execution_count": 26,
   "outputs": [
    {
     "output_type": "execute_result",
     "execution_count": 26,
     "data": {
      "text/plain": "12.615317765828324"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "trusted": true
   },
   "cell_type": "code",
   "source": "",
   "execution_count": null,
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3",
   "language": "python"
  },
  "language_info": {
   "name": "python",
   "version": "3.6.5",
   "mimetype": "text/x-python",
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "pygments_lexer": "ipython3",
   "nbconvert_exporter": "python",
   "file_extension": ".py"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}