Part 1A Sale Prediction via Basis Regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 1A: Sale Prediction via Basis Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Least Squares via Normal Equations, QR Decomposition & SVD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def lstsq_norm_e(A, b):\n",
    "    L = np.linalg.cholesky(np.dot(A.T, A))\n",
    "    R = L.T\n",
    "    A_tb = np.dot(A.T, b)\n",
    "    w = np.linalg.solve(R.T, A_tb)\n",
    "    return np.linalg.solve(R, w)\n",
    "\n",
    "\n",
    "def lstsq_qr(A, b):\n",
    "    Q, R = np.linalg.qr(A, mode='reduced')\n",
    "    Qtb = np.dot(Q.T, b)\n",
    "    return np.linalg.solve(R, Qtb)\n",
    "\n",
    "\n",
    "def lstsq_svd(A, b):\n",
    "    U,S,V = np.linalg.svd(A,full_matrices=False)\n",
    "    V = V.T\n",
    "    U_tb = np.dot(U.T,b)\n",
    "    S = np.diag(S)\n",
    "    w = np.linalg.solve(S, U_tb)\n",
    "    return np.dot(V,w)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Helper Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def p11_form_A(months):\n",
    "    N = months.shape[0]\n",
    "    A = np.zeros([N,8])\n",
    "    for i in range(N):\n",
    "        pi = np.pi\n",
    "        row = np.array([np.sin(pi * (1.0/12) * months[i]), np.cos(pi * (1.0/12) * months[i]),\n",
    "                np.sin(pi * (1.0/6) * months[i]), np.cos(pi * (1.0/6) * months[i]),\n",
    "                np.sin(pi * (1.0/2) * months[i]), np.cos(pi * (1.0/2) * months[i]),\n",
    "                months[i],\n",
    "                1])\n",
    "        A[i,:] = row\n",
    "    return A\n",
    "\n",
    "\n",
    "def p11_remove_nans(row_vector):\n",
    "    months = np.nonzero(~np.isnan(row_vector))[0]\n",
    "    sales = row_vector[months]\n",
    "    return months, sales"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Estimate Sales of Next 6 Months"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 1: Estimate\n",
    "def p11_basis_regression_all_items(sales_data):\n",
    "    num_item = sales_data.shape[0]\n",
    "    sales = np.zeros([num_item, 3, 6])  # which item, which method, which month?\n",
    "    months = np.zeros([num_item, 6])\n",
    "    for i in range(num_item):\n",
    "        months_current_item, sales_current_item = p11_remove_nans(sales_data[i, :])\n",
    "        future_months, future_sales = p11_basis_regression_one_item(months_current_item, sales_current_item)\n",
    "        print 'Item %d Predictions' % i\n",
    "        print future_sales, '\\n'\n",
    "        sales[i] = future_sales\n",
    "        months[i] = future_months\n",
    "    return months, sales\n",
    "\n",
    "\n",
    "def p11_basis_regression_one_item(months, sales):\n",
    "    A = p11_form_A(months)\n",
    "    b = sales\n",
    "\n",
    "    x_norm_e = lstsq_norm_e(A,b)\n",
    "    x_qr = lstsq_qr(A,b)\n",
    "    x_svd = lstsq_svd(A,b)\n",
    "\n",
    "    next_six_months = np.arange(months[-1]+1, months[-1]+7)\n",
    "    A_new = p11_form_A(next_six_months)\n",
    "\n",
    "    next_six_sales = np.zeros([3,6])  # 3 methods, 6 months estimated\n",
    "\n",
    "    next_six_sales[0,:] = np.dot(A_new, x_norm_e)\n",
    "    next_six_sales[1,:] = np.dot(A_new, x_qr)\n",
    "    next_six_sales[2,:] = np.dot(A_new, x_svd)\n",
    "    return next_six_months, next_six_sales"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Compare Different Methods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def p11_compare_methods_all_items(sales_estimated):\n",
    "    num_item = sales_estimated.shape[0]\n",
    "    comparison = np.zeros([num_item, 3, 3])  # which item, which method, compared to which method?\n",
    "    for i in range(num_item):\n",
    "        comparison_one_item = p11_compare_methods_one_item(sales_estimated[i, :, :])\n",
    "        print 'Item %d Comparison of Methods' % i\n",
    "        print comparison_one_item, '\\n'\n",
    "        comparison[i,:,:] = comparison_one_item\n",
    "    return comparison\n",
    "\n",
    "\n",
    "def p11_compare_methods_one_item(sales_estimated_for_item):  # (3,6) matrix\n",
    "    relative_error = np.zeros([3,3])\n",
    "    for i in range(3):\n",
    "        for j in range(3):\n",
    "            x = sales_estimated_for_item[i,:]\n",
    "            y = sales_estimated_for_item[j,:]\n",
    "            nom = np.linalg.norm(x - y)\n",
    "            denom = np.linalg.norm(y)\n",
    "            relative_error[i,j] = nom / denom\n",
    "    return relative_error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Find Error Compared to Ground Truth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def p11_gt_all_items(sales_estimated, test_data):\n",
    "    num_item = sales_estimated.shape[0]\n",
    "    gt_error = np.zeros([num_item, 3])  # which item, which method?\n",
    "    for i in range(num_item):\n",
    "        gt_error_one_item = p11_gt_one_item(sales_estimated[i, :], test_data[i, :])\n",
    "        print 'Item %d Error compared to Groud Truth' % i\n",
    "        print gt_error_one_item, '\\n'\n",
    "        gt_error[i,:] = gt_error_one_item\n",
    "    return gt_error\n",
    "\n",
    "\n",
    "def p11_gt_one_item(sales_estimated_for_item, test_data_for_item):  # (3,6) matrix\n",
    "    e = np.linalg.norm(sales_estimated_for_item - test_data_for_item, axis=1)\n",
    "    return e"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Call 3 Procedures Above, Plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def basis_regression(sales_data, test_data=None):\n",
    "    future_months, future_sales = p11_basis_regression_all_items(sales_data)\n",
    "    method_comparison = p11_compare_methods_all_items(future_sales)\n",
    "    gt_errors = None\n",
    "    if test_data is not None:\n",
    "        gt_errors = p11_gt_all_items(future_sales, test_data)\n",
    "    return future_months, future_sales, method_comparison, gt_errors\n",
    "\n",
    "\n",
    "def project_part1_1(train, test=None):\n",
    "    future_months, future_sales, comparison, gt_error = basis_regression(train, test)\n",
    "\n",
    "    plt.figure(figsize=(16, 8))\n",
    "    colors = ['r','k','g','b','m','y','c']\n",
    "    for i in range(train.shape[0]):\n",
    "        past_months, past_sales = p11_remove_nans(train[i, :])\n",
    "        plt.plot(past_months, past_sales, '--o', color=colors[i], lw=0.7, ms=6)\n",
    "        plt.plot(future_months[i,:], future_sales[i,0,:], '^', color=colors[i], ms=8)\n",
    "    plt.show()\n",
    "\n",
    "    errors = None\n",
    "    if test is not None:\n",
    "        errors = test - future_sales[:,0,:]\n",
    "\n",
    "    return future_months, future_sales, errors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test 1: Dataset without NaN's"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Item 0 Predictions\n",
      "[[ 383.07  386.7   404.29  392.67  360.95  354.16]\n",
      " [ 383.07  386.7   404.29  392.67  360.95  354.16]\n",
      " [ 383.07  386.7   404.29  392.67  360.95  354.16]] \n",
      "\n",
      "Item 1 Predictions\n",
      "[[ 575.82  559.44  582.84  592.78  567.52  561.97]\n",
      " [ 575.82  559.44  582.84  592.78  567.52  561.97]\n",
      " [ 575.82  559.44  582.84  592.78  567.52  561.97]] \n",
      "\n",
      "Item 2 Predictions\n",
      "[[ 1107.66  1067.86  1058.43  1071.17  1043.56   990.5 ]\n",
      " [ 1107.66  1067.86  1058.43  1071.17  1043.56   990.5 ]\n",
      " [ 1107.66  1067.86  1058.43  1071.17  1043.56   990.5 ]] \n",
      "\n",
      "Item 3 Predictions\n",
      "[[ 650.92  633.07  630.49  631.68  622.16  620.8 ]\n",
      " [ 650.92  633.07  630.49  631.68  622.16  620.8 ]\n",
      " [ 650.92  633.07  630.49  631.68  622.16  620.8 ]] \n",
      "\n",
      "Item 4 Predictions\n",
      "[[ 1008.84   997.1   1029.69  1026.92   974.4    957.51]\n",
      " [ 1008.84   997.1   1029.69  1026.92   974.4    957.51]\n",
      " [ 1008.84   997.1   1029.69  1026.92   974.4    957.51]] \n",
      "\n",
      "Item 0 Comparison of Methods\n",
      "[[  0.00e+00   5.94e-14   6.19e-14]\n",
      " [  5.94e-14   0.00e+00   2.57e-15]\n",
      " [  6.19e-14   2.57e-15   0.00e+00]] \n",
      "\n",
      "Item 1 Comparison of Methods\n",
      "[[  0.00e+00   9.56e-14   1.01e-13]\n",
      " [  9.56e-14   0.00e+00   5.37e-15]\n",
      " [  1.01e-13   5.37e-15   0.00e+00]] \n",
      "\n",
      "Item 2 Comparison of Methods\n",
      "[[  0.00e+00   9.36e-14   9.66e-14]\n",
      " [  9.36e-14   0.00e+00   3.14e-15]\n",
      " [  9.66e-14   3.14e-15   0.00e+00]] \n",
      "\n",
      "Item 3 Comparison of Methods\n",
      "[[  0.00e+00   1.65e-13   1.66e-13]\n",
      " [  1.65e-13   0.00e+00   1.35e-15]\n",
      " [  1.66e-13   1.35e-15   0.00e+00]] \n",
      "\n",
      "Item 4 Comparison of Methods\n",
      "[[  0.00e+00   4.85e-14   5.28e-14]\n",
      " [  4.85e-14   0.00e+00   4.39e-15]\n",
      " [  5.28e-14   4.39e-15   0.00e+00]] \n",
      "\n",
      "Item 0 Error compared to Groud Truth\n",
      "[ 47.32  47.32  47.32] \n",
      "\n",
      "Item 1 Error compared to Groud Truth\n",
      "[ 47.94  47.94  47.94] \n",
      "\n",
      "Item 2 Error compared to Groud Truth\n",
      "[ 100.05  100.05  100.05] \n",
      "\n",
      "Item 3 Error compared to Groud Truth\n",
      "[ 88.9  88.9  88.9] \n",
      "\n",
      "Item 4 Error compared to Groud Truth\n",
      "[ 86.32  86.32  86.32] \n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA64AAAHhCAYAAABul+PiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxFVERERERERsTYmriIiIiIiI2Nr/BxLMFzKDrzCmAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112075b10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Errors : \n",
      "[[  2.93  11.3   15.71  19.33  23.05  30.84]\n",
      " [ -0.82   8.56  13.16  20.22  23.48  33.03]\n",
      " [ 10.34  26.14  29.57  40.83  51.44  63.5 ]\n",
      " [  6.08  18.93  23.51  33.32  45.84  61.2 ]\n",
      " [  4.16  21.9   28.31  35.08  41.6   56.49]]\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "np.random.seed(12345)\n",
    "np.set_printoptions(precision=2,suppress=False)\n",
    "\n",
    "R_true = np.genfromtxt('dataset11.csv', delimiter=',',dtype=float)\n",
    "#train = R_true.copy()\n",
    "#test = None\n",
    "train = R_true[:,:18].copy()\n",
    "test = R_true[:,18:].copy()\n",
    "\n",
    "future_months, future_sales, errors = project_part1_1(train, test)\n",
    "\n",
    "print 'Errors : \\n', errors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test 2: Dataset with NaN's"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Item 0 Predictions\n",
      "[[ 1055.58  1088.39  1088.35  1067.2   1087.94  1135.92]\n",
      " [ 1055.58  1088.39  1088.35  1067.2   1087.94  1135.92]\n",
      " [ 1055.58  1088.39  1088.35  1067.2   1087.94  1135.92]] \n",
      "\n",
      "Item 1 Predictions\n",
      "[[ 714.21  751.31  772.22  790.35  816.14  829.59]\n",
      " [ 714.21  751.31  772.22  790.35  816.14  829.59]\n",
      " [ 714.21  751.31  772.22  790.35  816.14  829.59]] \n",
      "\n",
      "Item 2 Predictions\n",
      "[[ 399.14  387.29  359.63  360.02  378.63  371.55]\n",
      " [ 399.14  387.29  359.63  360.02  378.63  371.55]\n",
      " [ 399.14  387.29  359.63  360.02  378.63  371.55]] \n",
      "\n",
      "Item 3 Predictions\n",
      "[[ 633.52  653.65  631.04  620.95  644.22  645.99]\n",
      " [ 633.52  653.65  631.04  620.95  644.22  645.99]\n",
      " [ 633.52  653.65  631.04  620.95  644.22  645.99]] \n",
      "\n",
      "Item 4 Predictions\n",
      "[[ 1057.01  1072.63  1042.51  1048.69  1104.3   1123.99]\n",
      " [ 1057.01  1072.63  1042.51  1048.69  1104.3   1123.99]\n",
      " [ 1057.01  1072.63  1042.51  1048.69  1104.3   1123.99]] \n",
      "\n",
      "Item 0 Comparison of Methods\n",
      "[[  0.00e+00   2.56e-16   1.50e-15]\n",
      " [  2.56e-16   0.00e+00   1.72e-15]\n",
      " [  1.50e-15   1.72e-15   0.00e+00]] \n",
      "\n",
      "Item 1 Comparison of Methods\n",
      "[[  0.00e+00   2.51e-15   2.53e-15]\n",
      " [  2.51e-15   0.00e+00   2.85e-16]\n",
      " [  2.53e-15   2.85e-16   0.00e+00]] \n",
      "\n",
      "Item 2 Comparison of Methods\n",
      "[[  0.00e+00   1.29e-14   5.80e-14]\n",
      " [  1.29e-14   0.00e+00   4.52e-14]\n",
      " [  5.80e-14   4.52e-14   0.00e+00]] \n",
      "\n",
      "Item 3 Comparison of Methods\n",
      "[[  0.00e+00   4.62e-15   1.08e-14]\n",
      " [  4.62e-15   0.00e+00   1.54e-14]\n",
      " [  1.08e-14   1.54e-14   0.00e+00]] \n",
      "\n",
      "Item 4 Comparison of Methods\n",
      "[[  0.00e+00   1.22e-14   1.15e-14]\n",
      " [  1.22e-14   0.00e+00   2.07e-15]\n",
      " [  1.15e-14   2.07e-15   0.00e+00]] \n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA64AAAHhCAYAAABul+PiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncBUREREREZGwpsBVREREREREwtr/Aw75GsyokkxyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112102250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Errors : \n",
      "None\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "np.random.seed(12345)\n",
    "np.set_printoptions(precision=2,suppress=False)\n",
    "\n",
    "R_true = np.genfromtxt('dataset12.csv', delimiter=',',dtype=float)\n",
    "train = R_true.copy()\n",
    "test = None\n",
    "#train = R_true[:,:18].copy()\n",
    "#test = R_true[:,18:].copy()\n",
    "\n",
    "future_months, future_sales, errors = project_part1_1(train, test)\n",
    "\n",
    "print 'Errors : \\n', errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}