test-uxml.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import scipy.sparse as sp\n",
    "from scipy.sparse import csc_matrix as csc\n",
    "import pandas as pd\n",
    "pd.options.display.float_format = '{:,.6f}'.format\n",
    "from sklearn.metrics import (\n",
    "    mean_squared_error,\n",
    "    mean_absolute_error,\n",
    "    r2_score,\n",
    "    explained_variance_score,\n",
    "    roc_auc_score,\n",
    "    log_loss,\n",
    "    confusion_matrix\n",
    ")\n",
    "from sklearn.preprocessing import minmax_scale\n",
    "import time\n",
    "import matplotlib.pyplot as plt\n",
    "FONT_SIZE = 24\n",
    "plt.rcParams['figure.figsize'] = (20,8)\n",
    "plt.rcParams['font.size'] = FONT_SIZE\n",
    "plt.rcParams['legend.fontsize'] = FONT_SIZE\n",
    "plt.rcParams['xtick.labelsize'] = FONT_SIZE\n",
    "plt.rcParams['ytick.labelsize'] = FONT_SIZE\n",
    "%config InlineBackend.figure_format ='retina'\n",
    "import itertools\n",
    "from constants import (SEED, EPSILON, EVENT_THRESHOLD, DEFAULT_K, DEFAULT_THRESHOLD,  UX_CONSTANTS, LOG_DIR, \n",
    "                       DATA_DIR, TEST_DATA_PATH, DATA_OCT, DATA_NOV, USECOLS, USER, ITEM, RATING, PREDICTION)\n",
    "from utilities.ms_evaluation import (rmse, auc, logloss, precision_at_k, recall_at_k, ndcg_at_k, map_at_k, mae, rsquared, exp_var)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# NAME = r'BasicMatrixFactorization'\n",
    "NAME = r'AdamUXML' \n",
    "Y_HAT_PATH = DATA_DIR+r'/'+NAME+r'-y_hat.npz'\n",
    "TEST_RESULTS_PATH = LOG_DIR+'\\\\'+NAME+'\\\\test-results.csv'\n",
    "SKL = \"(calculated using sklearn.metrics on non-zero values of sparse matrices)\"\n",
    "SPA = \"(calculated using CSC sparse matrix operations)\"\n",
    "MSE = \"(calculated using the Microsoft Evaluation method)\"\n",
    "ux_constants = pd.Series(pd.read_csv(UX_CONSTANTS, index_col=0, squeeze=True, header=None), dtype='float32')\n",
    "POSITIVE_ABOVE = ux_constants['positive_above']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of the matrices: (177592, 44780)\n",
      "Number of non-zero values:\n",
      "Y:  552,255\n",
      "Ŷ:  552,255\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "log = pd.Series(dtype='float64')\n",
    "y_hat = sp.load_npz(Y_HAT_PATH) \n",
    "y = sp.load_npz(TEST_DATA_PATH)\n",
    "assert y_hat.shape == y.shape, 'The shape of Y and Y_hat must match, otherwise they are not comparable.'\n",
    "print(f\"Shape of the matrices: {y.shape}\")\n",
    "print(\"Number of non-zero values:\")\n",
    "print(f\"Y: {y.nnz:8,}\")\n",
    "print(f\"Ŷ: {y_hat.nnz:8,}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sparse_to_dense(matrix):\n",
    "    matrix = matrix.todok()\n",
    "    dense = []\n",
    "    rows,cols = matrix.nonzero()\n",
    "    for row,col in zip(rows,cols):\n",
    "        dense.append([row,col,matrix[row,col]])\n",
    "    return dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_true = pd.DataFrame(data=sparse_to_dense(y),columns=[USER, ITEM, RATING])\n",
    "df_pred = pd.DataFrame(data=sparse_to_dense(y_hat),columns=[USER, ITEM, PREDICTION])\n",
    "# Usually, the CSC is used when there are more rows than columns. (If there are more columns, use CSR instead.)\n",
    "y_hat = y_hat.tocsc()\n",
    "y = y.tocsc()\n",
    "y_nz = np.array(y[y.nonzero()]).reshape(-1)\n",
    "y_hat_nz = np.array(y_hat[y_hat.nonzero()]).reshape(-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Standard metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mean Square Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Square Error: 0.06813717876026473 (calculated using CSC sparse matrix operations)\n",
      "Mean Square Error: 0.06813716143369675 (calculated using sklearn.metrics on non-zero values of sparse matrices)\n",
      "Note: The smaller the better.\n"
     ]
    }
   ],
   "source": [
    "mse_spa = csc.sum(csc.power(y_hat-y,2))/y.nnz\n",
    "mse_skl = mean_squared_error(y_nz,y_hat_nz)\n",
    "print(f\"Mean Square Error: {mse_spa} {SPA}\")\n",
    "print(f\"Mean Square Error: {mse_skl} {SKL}\")\n",
    "print('Note: The smaller the better.')\n",
    "log[\"mse\"]=mse_spa"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Root Mean Square Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Root Mean Square Error: 0.26103099195357 (calculated using CSC sparse matrix operations)\n",
      "Root Mean Square Error: 0.26103097200393677 (calculated using sklearn.metrics on non-zero values of sparse matrices)\n",
      "Root Mean Square Error: 0.261030982048471 (calculated using the Microsoft Evaluation method)\n",
      "Note: The smaller the better.\n"
     ]
    }
   ],
   "source": [
    "rmse_spa = np.sqrt(mse_spa)\n",
    "rmse_skl = np.sqrt(mse_skl)\n",
    "rmse_mse = rmse(df_true,df_pred)\n",
    "print(f\"Root Mean Square Error: {rmse_spa} {SPA}\")\n",
    "print(f\"Root Mean Square Error: {rmse_skl} {SKL}\")\n",
    "print(f\"Root Mean Square Error: {rmse_mse} {MSE}\")\n",
    "print('Note: The smaller the better.')\n",
    "log[\"rmse\"]=rmse_spa"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mean Absolute Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Absolute Error: 0.17580741403427763 (calculated using CSC sparse matrix operations)\n",
      "Mean Absolute Error: 0.1758074015378952 (calculated using sklearn.metrics on non-zero values of sparse matrices)\n",
      "Mean Absolute Error: 0.17580742640347657 (calculated using the Microsoft Evaluation method)\n",
      "Note: The smaller the better.\n"
     ]
    }
   ],
   "source": [
    "mae_spa = csc.sum(abs(y_hat-y))/y.nnz\n",
    "mae_skl = mean_absolute_error(y_nz,y_hat_nz)\n",
    "mae_mse = mae(df_true,df_pred)\n",
    "print(f\"Mean Absolute Error: {mae_spa} {SPA}\")\n",
    "print(f\"Mean Absolute Error: {mae_skl} {SKL}\")\n",
    "print(f\"Mean Absolute Error: {mae_mse} {MSE}\")\n",
    "print('Note: The smaller the better.')\n",
    "log[\"mae\"]=mae_spa"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### R²"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Coefficient of determination (R²): 0.035397567799008156 (calculated using sklearn.metrics on non-zero values of sparse matrices)\n",
      "Coefficient of determination (R²): 0.03539762417084458 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "r2_skl = r2_score(y_nz,y_hat_nz)\n",
    "r2_mse = rsquared(df_true,df_pred)\n",
    "print(f\"Coefficient of determination (R\\u00B2): {r2_skl} {SKL}\")\n",
    "print(f\"Coefficient of determination (R\\u00B2): {r2_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[\"r-squared\"]=r2_skl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Explained variance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Explained variance: 0.03548389673233032 (calculated using sklearn.metrics on non-zero values of sparse matrices)\n",
      "Explained variance: 0.03548414320026483 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "exp_var_skl = explained_variance_score(y_nz,y_hat_nz)\n",
    "exp_var_mse = exp_var(df_true,df_pred)\n",
    "print(f\"Explained variance: {exp_var_skl} {SKL}\")\n",
    "print(f\"Explained variance: {exp_var_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[\"exp_var\"]=exp_var_skl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Arear Under Curve (AUC) - integral area under the receiver operating characteristic curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# For AUC and logloss we need binary ratings and predictions in the [0,1] range\n",
    "df_true_bin = df_true.copy()\n",
    "df_true_bin[RATING] = df_true_bin[RATING].apply(lambda x: 1 if x > POSITIVE_ABOVE else 0)\n",
    "df_pred_bin = df_pred.copy()\n",
    "df_pred_bin[PREDICTION] = minmax_scale(df_pred_bin[PREDICTION].astype(float))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Arear Under Curve (AUC): 0.8112665281656525 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better. 0.5 indicates an uninformative classifier\n"
     ]
    }
   ],
   "source": [
    "auc_mse = auc(df_true_bin,df_pred_bin)\n",
    "print(f\"Arear Under Curve (AUC): {auc_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better. 0.5 indicates an uninformative classifier\")\n",
    "log[\"auc\"]=auc_mse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic loss (logloss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic loss (logloss): 0.7089584715502268 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 0 the better.\n"
     ]
    }
   ],
   "source": [
    "logloss_mse = logloss(df_true_bin,df_pred_bin)\n",
    "print(f\"Logistic loss (logloss): {logloss_mse} {MSE}\")\n",
    "print(\"Note: The closer to 0 the better.\")\n",
    "log[\"logloss\"]=logloss_mse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 507,
       "width": 499
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "conf_mat = confusion_matrix(\n",
    "    [1 if x > POSITIVE_ABOVE  else 0 for x in y_nz],\n",
    "    [1 if x > POSITIVE_ABOVE  else 0 for x in y_hat_nz],  \n",
    "    labels=[1,0])\n",
    "\n",
    "conf_mat_p = np.round(conf_mat.astype('float') / conf_mat.sum(axis=1)[:, np.newaxis],4)*100\n",
    "log[\"true_positive\"]=f\"{conf_mat_p[0,0]:.2f}%\"\n",
    "log[\"false_negative\"]=f\"{conf_mat_p[0,1]:.2f}%\"\n",
    "log[\"false_positive\"]=f\"{conf_mat_p[1,0]:.2f}%\"\n",
    "log[\"true_negative\"]=f\"{conf_mat_p[1,1]:.2f}%\"\n",
    "\n",
    "precision = conf_mat[0,0]/(conf_mat[0,0]+conf_mat[1,0])\n",
    "recall = conf_mat[0,0]/(conf_mat[0,0]+conf_mat[0,1])\n",
    "log[\"precision\"]=precision\n",
    "log[\"recall\"]=recall\n",
    "log[\"F1\"] = (2*precision*recall)/(precision+recall)\n",
    "\n",
    "descriptions = np.array([[\"True Positive\", \"False Negative\"], [\"False Positive\", \"True Negatives\"]])\n",
    "colors = np.array([[\"green\", \"red\"], [\"red\", \"green\"]])\n",
    "\n",
    "plt.imshow([[0,0],[0,0]], interpolation='nearest', cmap=plt.cm.Greys)\n",
    "for i, j in itertools.product(range(conf_mat_p.shape[0]), range(conf_mat_p.shape[1])): \n",
    "    plt.text(j, i,\n",
    "             format(conf_mat[i, j], \",\")+'\\n'+\n",
    "             format(conf_mat_p[i, j], \".2f\")+'%\\n'+\n",
    "             descriptions[i, j], horizontalalignment=\"center\", verticalalignment=\"center\",\n",
    "             color=colors[i,j])\n",
    "plt.axhline(y=0.5, xmin=0, xmax=1, color=\"black\", linewidth=0.75)\n",
    "plt.axvline(x=0.5, ymin=0, ymax=1, color=\"black\", linewidth=0.75)\n",
    "plt.ylabel('True')\n",
    "plt.xlabel('Predicted')\n",
    "# plt.title(\"Confusion Matrix\")\n",
    "plt.xticks([0,1], [1,0])\n",
    "plt.yticks([0,1], [1,0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Precision @ K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision @ 10: 0.34191752705560396 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "precision_at_k_mse = precision_at_k(df_true,df_pred)\n",
    "print(f\"Precision @ {DEFAULT_K}: {precision_at_k_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[f\"precision-at-{DEFAULT_K}\"]=precision_at_k_mse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Recall @ K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recall @ 10: 0.9668412602660089 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "recall_at_k_mse = recall_at_k(df_true,df_pred)\n",
    "print(f\"Recall @ {DEFAULT_K}: {recall_at_k_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[f\"recall-at-{DEFAULT_K}\"]=recall_at_k_mse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### normalized Discounted Cumulative Gain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "normalized Discounted Cumulative Gain (nDCG@10): 1.0 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "ndcg_mse = ndcg_at_k(df_true,df_pred)\n",
    "print(f\"normalized Discounted Cumulative Gain (nDCG@{DEFAULT_K}): {ndcg_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[f\"ndcg-at-{DEFAULT_K}\"]=ndcg_mse"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### mAP (mean Average Precision) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean Average Precision (mAP@10): 0.9668412602660089 (calculated using the Microsoft Evaluation method)\n",
      "Note: The closer to 1 the better.\n"
     ]
    }
   ],
   "source": [
    "map_mse = map_at_k(df_true,df_pred)\n",
    "print(f\"mean Average Precision (mAP@{DEFAULT_K}): {map_mse} {MSE}\")\n",
    "print(\"Note: The closer to 1 the better.\")\n",
    "log[f\"map-at-{DEFAULT_K}\"]=map_mse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mse               0.068137\n",
       "rmse              0.261031\n",
       "mae               0.175807\n",
       "r-squared         0.035398\n",
       "exp_var           0.035484\n",
       "auc               0.811267\n",
       "logloss           0.708958\n",
       "true_positive       31.68%\n",
       "false_negative      68.32%\n",
       "false_positive       3.80%\n",
       "true_negative       96.20%\n",
       "precision         0.596072\n",
       "recall            0.316774\n",
       "F1                0.413696\n",
       "precision-at-10   0.341918\n",
       "recall-at-10      0.966841\n",
       "ndcg-at-10        1.000000\n",
       "map-at-10         0.966841\n",
       "dtype: object"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Elapsed time: 181.17 seconds\n"
     ]
    }
   ],
   "source": [
    "print(f\"Elapsed time: {time.time()-start_time:.2f} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Results saved to: C:\\TensorLogs\\AdamUXML\\test-results.csv\n"
     ]
    }
   ],
   "source": [
    "log.to_csv(TEST_RESULTS_PATH, index = True, header=False)\n",
    "print(f\"Results saved to: {TEST_RESULTS_PATH}\")"
   ]
  }
 ],
 "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"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}