{ "cells": [ { "cell_type": "code", "execution_count": 27, "metadata": { "collapsed": false }, "outputs": [], "source": [ "import numpy as np\n", "import scipy as sp\n", "import matplotlib.pyplot as plt\n", "import random\n", "from scipy import stats\n", "from scipy.optimize import fmin\n", "\n", "\n", "import mpld3\n", "mpld3.enable_notebook()\n", "figsize = (8,5)\n", "%matplotlib inline\n", "#figsize = (12, 8)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Quantile estimation via Stochastic gradient descent (Robbins Monro)\n", "\n", "Let $(X_n)_n$ be a sequence of i.i.d. real-valued random variables distributed like $X$. Let $(\\eta_n)_n$ be\n", "a non-increasing sequence of positive numbers (the step sizes). Finally, let\n", "$H:\\mathbb{R}^2\\mapsto\\mathbb{R}$ be a function. We consider the Robbins-Monro\n", "algorithm:\n", "\\begin{equation*}\n", " x_{k+1}=x_k-\\gamma_{k}H(x_k,X_{k+1})\n", "\\end{equation*}and $x_0\\in\\mathbb{R}$ is the starting point of the algorithm.\n", "\n", "Let $p\\in(0,1)$. For the quantile estimation problem via the Robbins-Monro algorithm,\n", "we consider the function $$H(x,X)=I( X\\leq x)-p.$$ In this\n", "particular case, the RM algorithm is\n", "\\begin{equation*}\n", " x_{k+1}=x_k-\\gamma_{k}\\big(I(X_{n+1}\\leq x_k)-p\\big) \\mbox{ and } x_0\\in\\mathbb{R}.\n", "\\end{equation*} \n", "We denote by $q_p$ the quantile of order $p$ of $X$:\n", "\\begin{equation*}\n", " q_p = \\min\\big(x\\in\\mathbb{R}:F(x)\\geq p\\big) \\mbox{ where }\n", " F(x)=\\mathbb{P}(X\\leq x).\n", "\\end{equation*}When $X$ has a cdf $F$ which is continuous and strictly increasing on some interval, $q_p$ is the unique solution to $F(x)=p$.\n", "\n", "The pseudo code of the algorithm stopped at step N is:\n", "\n", "    1:   Choose initial guess $x_0$
\n", "    2:   for $k=1,2,\\ldots, N$ do
\n", "    3:       $s_k=I(X_{k+1}\\leq x_k)-p$
\n", "    4:       $x_{k+1} = x_k - \\gamma_{k+1} s_k$
\n", "    5:   end for\n", "\n", "Thanks to SGD, we can make progress right away and continue to make progress as we go through the data set. Therefore, stochastic gradient descent is often preferred to gradient descent or here to the empirical quantile when dealing with large data sets.\n", "\n", "Unlike gradient descent, stochastic gradient descent will tend to oscillate near a minimum value rather than continuously getting closer. It may never actually converge to the minimum though. One way around this is to slowly decrease the step size $\\gamma_k$ like $1/k$ as the algorithm runs." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Simulating various densities" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "From now on, we focus on the median estimation (that is quantile estimation for $p=1/2$). We consider three different types of density such that: one has a cdf with a plateau of length 1 centered in $0$, one has a null density in the median and one having a density equal to $1$ at the median." ] }, { "cell_type": "code", "execution_count": 28, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe4AAAFCCAYAAAA3yT7DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "def f1(x):\n", " if -2" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "f, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize = figsize)\n", "x = np.linspace(-3,3,1000)\n", "y1 = [f1(ele) for ele in x]\n", "y2 = [f2(ele) for ele in x]\n", "y3 = [f3(ele) for ele in x]\n", "nb_sample = 250\n", "ax1.hist(S1(nb_sample), normed = True, bins=30)\n", "ax1.plot(x, y1, 'r-')\n", "ax1.set_title('f1')\n", "ax2.hist(S2(nb_sample),normed = True, bins=30)\n", "ax2.plot(x, y2, 'r-')\n", "ax2.set_title('f2')\n", "ax3.hist(S3(nb_sample),normed = True, bins=30)\n", "ax3.plot(x, y3, 'r-')\n", "ax3.set_title('f3')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Simulating Robbins Monro paths" ] }, { "cell_type": "code", "execution_count": 31, "metadata": { "collapsed": false }, "outputs": [], "source": [ "def RM(theta_0, horizon, data_nb, power_step=0.5):\n", " \"\"\" theta_0 = starting point; horizon : length of the path; data_nb = 1,2,3 for f1,f2,f3; power_step: power in step size\"\"\"\n", " if data_nb == 1:\n", " data = S1(horizon)\n", " if data_nb == 2:\n", " data = S2(horizon)\n", " if data_nb == 3:\n", " data = S3(horizon)\n", " theta = theta_0\n", " path = list()\n", " for i in range(horizon):\n", " gamma_k = 1./(i+1)**power_step # step size k^power_step\n", " if theta > data[i]:\n", " s_k = 0.5 # 1-p quand p=1/2 pour la médiane\n", " else:\n", " s_k = -0.5\n", " theta = theta - gamma_k*s_k\n", " #print data[i], theta\n", " path.append(theta)\n", " return path" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now, we plot the path towards the quantile, which is the median here ($=0$ for data_nb = $2$ or $3$ but which is the interval $[-1,1]$ for data_nb = $1$)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "###density $f_1$ with a plateau [-1,1] for the median" ] }, { "cell_type": "code", "execution_count": 32, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "x10000 = 0.136933289153 -- median = plateau [-1,1]\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEPCAYAAACKplkeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "horizon = 10000\n", "theta_0 = 2\n", "path = RM(theta_0, horizon, 1)\n", "print 'x{} = {} -- median = plateau [-1,1]'.format(horizon, path[horizon-1])\n", "#print len(iterations), len(path)\n", "plt.plot(range(horizon), path)\n", "plt.xlabel(\"iterations\")\n", "plt.ylabel(\"RM\")\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "distribution of the RM $x_n$ (1000 simulations of $x_{5000}$)" ] }, { "cell_type": "code", "execution_count": 33, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEACAYAAABWLgY0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAEjNJREFUeJzt3W2MXNd93/Hvr2JY1WkclnVAPRGQnJiwaTiF3VoWmqoe\nO7KgBImkQKglA7HZxDWCKs4TiiSiU0TrN4zsoM1DA79IWwlMEbNgW1eQEDsRvdUIEpxYcSzLkimG\nUhG2ohOu3Uhh07SGKenfF3tFjzfLfbgzszN79P0AC9577rn3/u8e8jd3zjwwVYUkqU1/Y9YFSJKm\nx5CXpIYZ8pLUMENekhpmyEtSwwx5SWrYmiGf5O4kS0meWNH+E0meSvJkko+MtB9M8nSSE0mun1bR\nkqSN2bHO9nuAfwP81ssNSd4B3Ah8d1WdS/IdXft+4FZgP3A58Okk+6rqpalULkla15p38lX1MPD8\niuZ/DvxSVZ3r+ny1a78JOFJV56rqFPAMcPVky5UkbUafOfnXAf84yR8kGSb5B137ZcDpkX6nWb6j\nlyTNyHrTNRfa5+9U1TVJ3gocBV57gb5+Z4IkzVCfkD8NfAKgqv4wyUtJXgN8Gdg70u+Kru2bJDH4\nJamHqspm9+kzXXMv8E6AJPuAnVX1v4D7gNuS7ExyFcvTOo9eoNBmf+68886Z1+D1eX2vxOtr+dqq\n+t8br3knn+QI8Hbg7yZ5FvhF4G7g7u5tlV8H3tcF9/EkR4HjwAvA7TVOZY1LNv2A3ItDIL2yrRny\nVfWeC2x67wX6HwIOjVvUK8e0A3hrHkgkzS8/8Tphg8Fg1iVMlde3vbV8fS1f2ziy1U/nkziLw8vT\nNdO/k/d3LbUhCbVFL7xKkrYJQ16SGmbIS1LDDHlJapghL0kNM+QlqWGGvCQ1zJCXpIYZ8pLUMENe\nkhpmyEtSwwx5SWqYIS9JDTPkJalhhrwkNcyQl6SGGfKS1LA1Qz7J3UmWuv+0e+W2f5HkpSS7R9oO\nJnk6yYkk10+jYEnSxq13J38PcMPKxiR7gXcB/2OkbT9wK7C/2+djSXymIEkztGYIV9XDwPOrbPrX\nwM+taLsJOFJV56rqFPAMcPUkipQk9bPpO+0kNwGnq+qLKzZdBpweWT8NXD5GbZKkMe3YTOckrwI+\nxPJUzfnmNXapPkVJkiZjUyEPfCdwJfB4EoArgD9K8jbgy8Dekb5XdG1/zcLCwvnlwWDAYDDYZBmS\n1LbhcMhwOBz7OKla+2Y7yZXA/VX1plW2/Qnw96vque6F14+zPA9/OfBp4LtqxQmSrGx6RVp+kJz2\n7yH4u5bakISqWmvmZFXrvYXyCPAZYF+SZ5P8yIou5xOkqo4DR4HjwKeA201zSZqtde/kJ35C7+QB\n7+Qlbc5U7uQlSdubIS9JDTPkJalhhrwkNcyQl6SGGfKS1DBDXpIaZshLUsMMeUlqmCEvSQ0z5CWp\nYYa8JDXMkJekhhnyktQwQ16SGmbIS1LDDHlJapghL0kNM+QlqWHr/UfedydZSvLESNsvJ3kqyeNJ\nPpHk20e2HUzydJITSa6fZuGSpPWtdyd/D3DDirYHgDdW1d8DTgIHAZLsB24F9nf7fCyJzxQkaYbW\nDOGqehh4fkXbsap6qVv9LHBFt3wTcKSqzlXVKeAZ4OrJlitJ2oxx77R/FPhkt3wZcHpk22ng8jGP\nL0kaw46+Oyb5BeDrVfXxNbrVao0LCwvnlweDAYPBoG8ZktSk4XDIcDgc+zipWjWHv9EhuRK4v6re\nNNL2T4EPAN9bVV/r2u4AqKq7uvXfBe6sqs+uOF6td85XgiRc4DFwkmfB37XUhiRUVTa736ana5Lc\nAPwscNPLAd+5D7gtyc4kVwGvAx7d7PE1WUmm/iNpfq05XZPkCPB24DVJngXuZPndNDuBY90/8N+v\nqtur6niSo8Bx4AXgdm/Z58H0ny1Iml/rTtdM/IRO1wBbN13jlJDUhr7TNb1feJVethVTNj6QSP0Y\n8poAp4SkeeUnUiWpYYa8JDXM6ZoVlpaWeOihh2ZdhiRNhCG/wpNPPsn73vdBdu4cTO0cL7741akd\nW5JGGfKruPjiN3H27NEpnuER4NopHl+SljknL0kNM+QlqWGGvCQ1zJCXpIYZ8pLUMENekhpmyEtS\nwwx5SWqYH4bStuDXGUv9GPLaJvw6Y6kPp2skqWGGvCQ1bM2QT3J3kqUkT4y07U5yLMnJJA8k2TWy\n7WCSp5OcSHL9NAuXJK1vvTv5e4AbVrTdARyrqn3AYrdOkv3ArcD+bp+PJfGZgiTN0JohXFUPA8+v\naL4RONwtHwZu7pZvAo5U1bmqOgU8A1w9uVIlSZvV5057T1UtdctLwJ5u+TLg9Ei/08DlY9QmSRrT\nWG+hrKpKstZ721bdtrCwcH55MBgwGAzGKUOaCN+Lr3kyHA4ZDodjH6dPyC8luaSqziS5FPhK1/5l\nYO9Ivyu6tr9mNOSl+eF78TU/Vt4Af/jDH+51nD7TNfcBB7rlA8C9I+23JdmZ5CrgdcCjvaqSJE3E\nmnfySY4Abwdek+RZ4BeBu4CjSd4PnALeDVBVx5McBY4DLwC3l89NJWmm1gz5qnrPBTZdd4H+h4BD\n4xYlSZoM38cuSQ0z5CWpYYa8JDXMkJekhhnyktQwQ16SGmbIS1LDDHlJapghL0kNM+QlqWGGvCQ1\nzJCXpIYZ8pLUMENekhpmyEtSwwx5SWqYIS9JDTPkJalhhrwkNax3yCc5mORLSZ5I8vEkfzPJ7iTH\nkpxM8kCSXZMsVpK0Ob1CPsmVwAeAt1TVm4CLgNuAO4BjVbUPWOzWJUkz0vdO/n8D54BXJdkBvAr4\nU+BG4HDX5zBw89gVSpJ66xXyVfUc8K+A/8lyuP9FVR0D9lTVUtdtCdgzkSolSb3s6LNTku8Efhq4\nEjgL/KckPzzap6oqSa22/8LCwvnlwWDAYDDoU4YkNWs4HDIcDsc+TqpWzeG1d0puBd5VVf+sW38v\ncA3wTuAdVXUmyaXAg1X1+hX7Vp9zbpXFxUVuueUQZ88uTvEsjwDXAtP+PcRzzNk55vnvvuZbEqoq\nm92v75z8CeCaJH8rSYDrgOPA/cCBrs8B4N6ex5ealGTqP9KoXtM1VfV4kt8CPge8BHwe+E3g24Cj\nSd4PnALePaE6pUZsxTMS6Rt6hTxAVX0U+OiK5udYvquX1Kiterbg1NZk9A55Sa9kPiPZLvxaA0lq\nmCEvSQ0z5CWpYYa8JDXMkJekhhnyktQwQ16SGmbIS1LDDHlJapghL0kNM+QlqWGGvCQ1zC8okxrj\nd8prlCEvNcdviNQ3OF0jSQ0z5CWpYYa8JDXMOXlJc8n/ZnAyet/JJ9mV5D8neSrJ8SRvS7I7ybEk\nJ5M8kGTXJIuV9EpSW/DTvnGma34N+GRVvQH4buAEcAdwrKr2AYvduiRpRnqFfJJvB66tqrsBquqF\nqjoL3Agc7rodBm6eSJWSpF763slfBXw1yT1JPp/k3yb5VmBPVS11fZaAPROpUpLUS98XXncAbwE+\nWFV/mORXWTE1U1WVZNVJr4WFhfPLg8GAwWDQswxJatNwOGQ4HI59nPR5ZTnJJcDvV9VV3fo/Ag4C\nrwXeUVVnklwKPFhVr1+xb83zq9mLi4vccsshzp5dnOJZHgGuZWs+meg5PIfnWOs885xHo5JQVZt+\ny1Gv6ZqqOgM8m2Rf13Qd8CXgfuBA13YAuLfP8SVJkzHO++R/AvjtJDuB/w78CHARcDTJ+4FTwLvH\nrlCS1FvvkK+qx4G3rrLpuv7lSJImya81kKSGGfKS1DBDXpIaZshLUsMMeUlqmCEvSQ0z5CWpYYa8\nJDXMkJekhhnyktQwQ16SGmbIS1LDDHlJapghL0kNM+QlqWGGvCQ1zJCXpIYZ8pLUMENekho2Vsgn\nuSjJY0nu79Z3JzmW5GSSB5LsmkyZkqQ+xr2T/yngOFDd+h3AsaraByx265KkGekd8kmuAL4f+HdA\nuuYbgcPd8mHg5rGqkySNZZw7+V8BfhZ4aaRtT1UtdctLwJ4xji9JGtOOPjsl+QHgK1X1WJLBan2q\nqpLUatsWFhbOLw8GAwaDVQ8hSa9Yw+GQ4XA49nFStWoOr71Tcgh4L/ACcDHwauATwFuBQVWdSXIp\n8GBVvX7FvtXnnFtlcXGRW245xNmzi1M8yyPAtXzjpYxpiefwHJ5jnfPMcx6NSkJVZf2e36zXdE1V\nfaiq9lbVVcBtwH+rqvcC9wEHum4HgHv7HF+SNBmTep/8yw+FdwHvSnISeGe3LkmakV5z8qOq6iHg\noW75OeC6cY8pSZoMP/EqSQ0z5CWpYYa8JDXMkJekhhnyktQwQ16SGmbIS1LDDHlJapghL0kNM+Ql\nqWGGvCQ1zJCXpIYZ8pLUMENekhpmyEtSwwx5SWqYIS9JDTPkJalhvUI+yd4kDyb5UpInk/xk1747\nybEkJ5M8kGTXZMuVJG1G3zv5c8DPVNUbgWuAH0/yBuAO4FhV7QMWu3VJ0oz0CvmqOlNVX+iW/w/w\nFHA5cCNwuOt2GLh5EkVKkvoZe04+yZXAm4HPAnuqaqnbtATsGff4kqT+xgr5JH8b+C/AT1XVX45u\nq6oCapzjS5LGs6Pvjkm+heWA/w9VdW/XvJTkkqo6k+RS4Cur7buwsHB+eTAYMBgM+pYhSU0aDocM\nh8Oxj5PlG+5N7pSE5Tn3P6+qnxlp/2jX9pEkdwC7quqOFftWn3NulcXFRW655RBnzy5O8SyPANcy\n/Sc68Ryew3Osc555zqNRSaiqbHa/vnfy3wP8MPDFJI91bQeBu4CjSd4PnALe3fP4kqQJ6BXyVfUI\nF57Pv65/OZKkSfITr5LUMENekhpmyEtSwwx5SWqYIS9JDTPkJalhhrwkNcyQl6SGGfKS1DBDXpIa\nZshLUsMMeUlqmCEvSQ0z5CWpYYa8JDXMkJekhhnyktQwQ16SGmbIS1LDJh7ySW5IciLJ00l+ftLH\nlyRt3ERDPslFwG8ANwD7gfckecMkzzH/hrMuYMqGsy5gyoazLmDKhrMuYIqGsy5gLk36Tv5q4Jmq\nOlVV54D/CNw04XPMueGsC5iy4awLmLLhrAuYsuGsC5ii4awLmEuTDvnLgWdH1k93bZKkGdgx4ePV\nhI83E1/72uO8+tU/2HPfP+bii/9ozT4vvvjn/NVf9Tq8JG1KqiaXy0muARaq6oZu/SDwUlV9ZKRP\nEw8EkrTVqiqb3WfSIb8D+GPge4E/BR4F3lNVT03sJJKkDZvodE1VvZDkg8DvARcB/96Al6TZmeid\nvCRpvkz9E69J/kmSLyV5Mclb1uh3KskXkzyW5NFp1zUpm7i+bfkhsSS7kxxLcjLJA0l2XaDfthq/\njYxHkl/vtj+e5M1bXWNf611bkkGSs91YPZbkX86izj6S3J1kKckTa/TZluMG619fr7Grqqn+AK8H\n9gEPAm9Zo9+fALunXc8sro/lqatngCuBbwG+ALxh1rVv8Po+Cvxct/zzwF3bffw2Mh7A9wOf7Jbf\nBvzBrOue4LUNgPtmXWvP67sWeDPwxAW2b8tx28T1bXrspn4nX1UnqurkBrtv+pXjWdvg9W3nD4nd\nCBzulg8DN6/Rd7uM30bG4/x1V9VngV1J9mxtmb1s9O/adhmrb1JVDwPPr9Flu44bsKHrg02O3Tx9\nQVkBn07yuSQfmHUxE7adPyS2p6qWuuUl4EL/YLbT+G1kPFbrc8WU65qEjVxbAf+wm874ZJL9W1bd\n9G3XcduoTY/dRN5dk+QYcMkqmz5UVfdv8DDfU1V/luQ7gGNJTnSPajM3geub61e317i+Xxhdqapa\n43MOczt+q9joeKy8Y5rrcexspMbPA3ur6v8m+T7gXpanHFuxHcdtozY9dhMJ+ap61wSO8Wfdn19N\n8l9Zfto5FyExgev7MrB3ZH0vy3cYc2Gt6+teBLqkqs4kuRT4ygWOMbfjt4qNjMfKPld0bfNu3Wur\nqr8cWf5Uko8l2V1Vz21RjdO0XcdtQ/qM3VZP16w6l5TkVUm+rVv+VuB64IKvns+xC82VfQ54XZIr\nk+wEbgXu27qyxnIfcKBbPsDyncM32Ybjt5HxuA94H5z/JPdfjExbzbN1ry3JniTplq9m+a3ULQQ8\nbN9x25BeY7cFrxb/EMtzZP8POAN8qmu/DPidbvm1LL8L4AvAk8DBWb/KPcnr69a/j+VPAz+zza5v\nN/Bp4CTwALCrhfFbbTyAHwN+bKTPb3TbH2eNd4bN28961wb8eDdOXwA+A1wz65o3cW1HWP40/de7\nf3c/2sq4beT6+oydH4aSpIbN07trJEkTZshLUsMMeUlqmCEvSQ0z5CWpYYa8JDXMkJekhhnyktSw\n/w+PK5PgPLO0WQAAAABJRU5ErkJggg==\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "list_theta_limit = list()\n", "horizon = 5000\n", "theta_0=-2\n", "for i in range(1000):\n", " path = RM(theta_0, horizon, 1)\n", " list_theta_limit.append(path[horizon-1])\n", "\n", "plt.hist(list_theta_limit)\n", "plt.show() " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For the data set 1, it looks like the limit of the RM algo is either [uniformly distributed over the plateau] or [the limiting law depends on the starting point (if theta_0 is >0 or <0)]. It is not clear yet. Moreover, it really looks like the limiting depends on the step size and in particular, how fast the serie $\\sum_k \\gamma_k$ goes to $\\infty$." ] }, { "cell_type": "markdown", "metadata": { "collapsed": false }, "source": [ "###density $f_2$ with no plateau for the median but for which f(q_{1/2})=0\n", "For the data set number 2 (for which f(q)=0 but for which there is no plateau), the step size gamma_k=1/k does not lead to the convergence of the algorithm. It works for gamma_k=1/sqrt(k). " ] }, { "cell_type": "code", "execution_count": 34, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "x10000 = -0.136316581064 -- median = 0\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEPCAYAAACKplkeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEZE1DCpERGQNgwoREVnDoEJERNb8P/JQgSCvzTj2AAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "horizon = 10000\n", "theta_0 = 2\n", "path = RM(theta_0, horizon, 2)\n", "print 'x{} = {} -- median = 0'.format(horizon, path[horizon-1])\n", "#print len(iterations), len(path)\n", "plt.plot(range(horizon), path)\n", "plt.xlabel(\"iterations\")\n", "plt.ylabel(\"RM\")\n", "plt.plot(range(horizon), np.zeros(horizon),color = 'r')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "distribution of the RM $x_n$ (1000 simulations of $x_{5000}$) -- when $f(q_p)=0$" ] }, { "cell_type": "code", "execution_count": 35, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n6T/uUy5NFAAAAABJRU5ErkJggg==\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "list_theta_limit = list()\n", "horizon = 5000\n", "theta_0=-2\n", "for i in range(1000):\n", " path = RM(theta_0, horizon, 2)\n", " list_theta_limit.append(path[horizon-1])\n", "\n", "plt.hist(list_theta_limit)\n", "plt.show() " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "###density $f_3$ for which the median $=0$ and $f(0)=1$" ] }, { "cell_type": "code", "execution_count": 36, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "x10000 = 0.00381726226498 -- median = 0\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEPCAYAAACKplkeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZoVxqJiZWWH+P4tH+eJYqbx5AAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "horizon = 10000\n", "theta_0 = 2\n", "path = RM(theta_0, horizon, 3, power_step=1)\n", "print 'x{} = {} -- median = 0'.format(horizon, path[horizon-1])\n", "#print len(iterations), len(path)\n", "plt.plot(range(horizon), path)\n", "plt.xlabel(\"iterations\")\n", "plt.ylabel(\"RM\")\n", "plt.plot(range(horizon), np.zeros(horizon),color = 'r')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "distribution of the RM $x_n$ (1000 simulations of $x_{5000}$): when $f(q_p)>1/2$,\n", "$$\\sqrt{n}(x_n-q_p)\\stackrel{d}{\\rightarrow} \\mathcal{N}(0,\\sigma^2) \\mbox{ where } \\sigma^2=\\frac{p(1-p)}{2f(q_p)-1}$$\n", "Here : $p=1/2$, $q_p=0$, $f(q_p)=1$ and $\\sigma^2=1/4$ and so\n", "$$x_n\\approx \\mathcal{N}(0,1/(4n))$$" ] }, { "cell_type": "code", "execution_count": 37, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeYAAAE4CAYAAACOp1CQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from scipy.stats import norm\n", "list_theta_limit = list()\n", "horizon = 5000\n", "theta_0=-2\n", "nb_paths = 1000\n", "for i in range(nb_paths):\n", " path = RM(theta_0, horizon, 3, power_step=1)\n", " list_theta_limit.append(path[horizon-1])\n", "####Gaussian limit\n", "Fx = 0\n", "var = 1/float(4)\n", "x = np.linspace(min(list_theta_limit), max(list_theta_limit), 1000)\n", "y = norm.pdf(x, loc = Fx, scale = np.sqrt(var/float(horizon)))\n", "####Plot\n", "f, ax = plt.subplots(1, 1, figsize = figsize)\n", "ax.hist(list_theta_limit, normed=True, bins=50)\n", "ax.plot(x,y)\n", "plt.show() " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##Comparison of estimators $x_n$ (RM) and empirical quantile $\\widehat q_{n,p}$ for density $f_3$" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "###limit law of empirical quantile\n", "$$\\sqrt{n}(\\widehat q_{n,p}-q_p)\\stackrel{d}{\\rightarrow}\\mathcal{N}(0,\\sigma^2)$$\n", "where $\\sigma^2= p(1-p)/f(q_p)^2$. \n", " \n", "Here: $p=1/2$, $q_p=0$, $f(q_p)=1$ and $\\sigma^2=1/4$\n" ] }, { "cell_type": "code", "execution_count": 19, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "horizon = 500, number of paths=1000\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeYAAAE4CAYAAACOp1CQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/html": [ "\n", "\n", "\n", "\n", "
\n", "" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from scipy.stats import norm\n", "#paramètres\n", "nb_paths, horizon, bins = 1000, 500, 20\n", "print 'horizon = {}, number of paths={}'.format(horizon, nb_paths)\n", "#data\n", "Hatq = list()\n", "for j in range(nb_paths):\n", " data0 = S3(horizon)\n", " Hatq.append(np.median(data0))\n", "#Gaussienne\n", "Fx = 0\n", "var = 1/float(4)\n", "x = np.linspace(min(Hatq), max(Hatq), 1000)\n", "y = norm.pdf(x, loc = Fx, scale = np.sqrt(var/float(horizon)))\n", "#plot\n", "f, ax = plt.subplots(1,1,figsize = figsize)\n", "ax.hist(Hatq, bins = 50, normed = True)\n", "ax.plot(x, y, 'r')\n", "#ax.set_title('x = {}, var = {}'.format(x, var))\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "###Comparing the speed of convergence of $\\widehat q_{n,p}$ and $x_n$\n", "Here, for the median of $f_3$ both $\\widehat q_{n,p}$ and $x_n$ converge towards $q_p=0$ at the speed of $1/\\sqrt{n}$ with the same asymptotic variance = 1/4. " ] }, { "cell_type": "code", "execution_count": 20, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "horizon = 10000 \n", " quantile = 0.00577366960229\n", " RM = 0.00534791670922\n", " theoretical median = 0\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "/Users/lecueguillaume/anaconda/lib/python2.7/site-packages/numpy/core/_methods.py:59: RuntimeWarning: Mean of empty slice.\n", " warnings.warn(\"Mean of empty slice.\", RuntimeWarning)\n", "/Users/lecueguillaume/anaconda/lib/python2.7/site-packages/numpy/core/_methods.py:71: RuntimeWarning: invalid value encountered in double_scalars\n", " ret = ret.dtype.type(ret / rcount)\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEPCAYAAACKplkeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xd8XNWZ//HPM6Nuyb0J4xZwoRoImBKKKCEOWUxJADuh\nhGSzJPwIpGwWyCYgSF4BNptsNvml8AtsQkkwJQRMCS2LIBCwKcY0G2PAjnuX1cvMPL8/7pU9kiV5\nxprRWNL3/XrNS7fNPc8czDxzz7nnXHN3REREMiGS6wBERKT/UFIREZGMUVIREZGMUVIREZGMUVIR\nEZGMUVIREZGMyWlSMbNZZrbUzN43s6u7OKbCzBaZ2dtmVtXLIYqISBosV+NUzCwKvAecBqwBXgHm\nuvuSpGOGAi8Cn3L31WY20t035yRgERHZrVxeqcwElrv7CndvBeYBZ3U45vPAn9x9NYASiojI3i2X\nSWUcsCppfXW4LdkUYLiZPWtmr5rZRb0WnYiIpC0vh2Wn0u6WDxwBnAqUAC+Z2cvu/n5WIxMRkT2S\ny6SyBhiftD6e4Gol2Spgs7s3Ao1m9jwwA2iXVMxME5iJiKTJ3S3T58xl89erwBQzm2RmBcAFwPwO\nxzwMHG9mUTMrAY4G3u3sZO4+4F/XX399zmPYW16qC9WF6qL7V7bk7ErF3WNmdgXwJBAFbnf3JWZ2\nWbj/VndfamZPAG8CCeC37t5pUhERkdzLZfMX7v4X4C8dtt3aYf0/gf/szbhERGTPaER9P1JRUZHr\nEPYaqoudVBc7qS6yL2eDHzPJzLw/fA4Rkd5iZngWOupz2vyVbWYZry/phhK7iPTrpAL6oustSuAi\nAupTERGRDFJSERGRjFFSERGRjFFSGeDOOOMM7rrrLgB+//vfc8IJJ+Q4IhHpy/p9R73sVFlZyQcf\nfLAjiQA8/vjjOYxIRPobXamIiEjGKKnkyKJFizjiiCMYPHgwc+bMYc6cOXz/+9/vtAkqEonw4Ycf\nAvDYY49x+OGHM2TIECZMmMANN9yw47gVK1YQiUS48847mThxIqNGjeJHP/oRAE888QQ33XQT9957\nL2VlZRx++OFAMML49ttv7zTGpUuX8slPfpIRI0Ywffp07r///mxUhYj0I0oqOdDS0sLZZ5/NJZdc\nwrZt2zjvvPN48MEHMbPdjvcoLS3l7rvvZvv27Tz22GP8+te/5uGHH253zIsvvsiyZcv461//yo03\n3sh7773HrFmz+O53v8ucOXOora1l0aJFAF2WWV9fzyc/+UkuvPBCNm3axLx587j88stZsmTJLseK\niLQZ0EnFLDOvdL388svEYjGuuuoqotEon/3sZznqqKNSGqh50kkncdBBBwFwyCGHMGfOHJ577rl2\nx1x//fUUFhZy6KGHMmPGDBYvXgyQ1pTXjz76KJMnT+aSSy4hEolw2GGHce655+pqRUS6NaA76nM1\n2H7t2rWMG9f+yckTJ05M6b0LFizgmmuu4Z133qGlpYXm5mbOP//8dseMHTt2x3JJSQl1dXVpx7hy\n5UoWLFjAsGHDdmyLxWJcfPHFaZ9LRAaOAX2lkivl5eWsWbOm3baVK1cCMGjQIBoaGnZsX79+fbvj\nPv/5z3P22WezevVqqqur+epXv0oikUip3HSmUpkwYQInnXQS27Zt2/Gqra3ll7/8ZcrnEJGBR0kl\nB4477jjy8vL4+c9/TmtrKw8++CCvvPIKZsaMGTN45513WLx4MU1NTVRWVrZ7b11dHcOGDaOgoICF\nCxfyxz/+MeVkMXbsWFasWJFSE9hnPvMZli1bxt13301rayutra288sorLF26dE8+sogMEEoqOZCf\nn8+DDz7I73//e0aMGMF9993Hueeei7szZcoUrrvuOk477TSmTZvGCSec0C5p/OpXv+K6665j8ODB\n/OAHP+CCCy5od+7uEsx5550HwIgRIzjyyCN32Z/caV9WVsZTTz3FvHnzGDduHOXl5Vx77bW0tLRk\nogpEpJ/q189TCZ8XkIOI0nfppZey77778oMf/CDXoeyRvlTXIpK956noSmUvoS9kEekPcppUzGyW\nmS01s/fN7OpO9leY2XYzWxS+vpeLOHtDKmNURET2djlr/jKzKPAecBqwBngFmOvuS5KOqQC+5e6z\nd3OuPt/81deprkX6lv7Y/DUTWO7uK9y9FZgHnNXJcfr5LiLSR+QyqYwDViWtrw63JXPgODNbbGaP\nm9mBvRadiIikLZcj6lNpK3kdGO/uDWb2aeAhYGp2wxIRkT2Vy6SyBhiftD6e4GplB3evTVr+i5n9\nysyGu/vWjidLHiRYUVFBRUVFpuMVEemzqqqqqKqqyno5ueyozyPoqD8VWAssZNeO+jHARnd3M5sJ\n3Ofukzo5lzrqc0x1LdK3ZKujPmdXKu4eM7MrgCeBKHC7uy8xs8vC/bcCnwO+ZmYxoAGYk6t4RURk\n93I6TsXd/+Lu09x9f3e/Kdx2a5hQcPdfuvvB7n6Yux/n7i/nMt5MmTRpEiUlJZSVlTF27Fguuugi\nampqAPjiF79IJBJh/vz57d7zzW9+k0gkwh133JGLkEVEUqIR9TlgZjz66KPU1tayePFi3nrrLX74\nwx/u2D916lTuvPPOHeuxWIz77ruP/fffXwMkRWSvpqSSY2PGjOH000/n3XffBYKEc+aZZ/LCCy9Q\nXV0NBI8CnjFjBmPGjFG/hYjs1ZRUcqQtOaxevZonnniCmTNn7thXVFTEWWedxbx58wC48847dzwc\nS1cqIrI3G9BPfrQbMvMF7dend/Xg7px99tmYGXV1dcyePZvvfa/9tGYXX3wx3/nOd5g7dy7PP/88\nd955px6QJSJ7vQGdVNJNBpliZjz88MOccsopPP/885x55pm8+uqrO65WzIxPfOITbNq0iR/+8Iec\neeaZFBUV5SRWEZF0qPkrx0488US+/vWvc/XVu0zSzIUXXshPf/pTPRdeRPoMJZW9wDe+8Q0WLlzI\nggULgJ39LVdeeSXPPPMMJ5xwQi7DExFJmZLKXmDkyJFccskl3HzzzcDOzvhhw4Zx8skn5zI0EZG0\n6HHCkhGqa5G+pT8+T0VERPoZJRUREckYJRUREckYJRUREckYJRUREckYJRUREcmY3U7TYmYR4AvA\nZHe/0cwmAGPdfWHWo8sATcAoItJ7djtOxcx+AySAU9x9upkNB55y9yN7I8BUdDVORUREOpfLxwkf\n7e6Hm9kiAHffamb5mQ5ERET6vlT6VFrMLNq2YmajCK5cRERE2kklqfwC+DMw2sx+BLwI3JSJws1s\nlpktNbP3zWzXaXp3HneUmcXM7NxMlCsiItmR0txfZnYAcGq4+ld3X9LjgoOrn/eA04A1wCvA3I7n\nDo97GmgAfufuf+rkXOpTERFJQ6/3qYQd8m02APeEy25mw919aw/Lngksd/cVYXnzgLOAjgnr68AD\nwFE9LE9ERLKsu47614Hufv5P7mHZ44BVSeurgaOTDzCzcQSJ5hSCpKLLERGRvViXScXdJ2W57FQS\nxM+Aa9zdLRhw0uWlWmVl5Y7liooKKioqehqfiEi/UVVVRVVVVdbL6bJPxcymu/tSMzuis/3u/nqP\nCjY7Bqh091nh+rVAwt1vSTrmQ3YmkpEE/Spfcff5Hc6lPhURkTRkq0+lu6TyW3f/iplV0clVhbv3\n6JGEZpZH0FF/KrAWWEgnHfVJx/8OeMTdH+xkn5KKiEgaer2j3t2/Ei7OcvemDsEU9bRgd4+Z2RXA\nk0AUuN3dl5jZZeH+W3tahoiI9K5Upml53d2P2N22XNKViohIenJxS3E5sA9QEvarGEEz2GCgJNOB\niIhI39fdLcWnA18kuPX3J0nba4HvZjEmERHpo1Jp/vqcuz/QS/HsETV/iYikp9fv/koquAj4LDCJ\noEPdAHf3GzMdzJ5SUhERSU8up75/GKgGXgOadnOsiIgMYKkklXHu/qmsRyIiIn1eKlPf/93MDs16\nJCIi0uel0qeyBNgf+AhoDje7u+81iUZ9KiIi6clln8qnM12oiIj0T7tNKknPOxkN9Hh6FhER6b92\n26diZrPN7H2C5q/ngBXAX7Icl4iI9EGpdNT/EDgWWObukwlmFV6Q1ahERKRPSiWptLr7ZiBiZlF3\nfxY4MstxiYhIH5RKR/02MysD/gb8wcw2AnXZDUtERPqiVG4pLgUaCa5qvkAwS/Ef3H1L9sNLjW4p\nFhFJT87m/uoLlFRERNKTs3EqZlbHzscJFwD5QJ27D850MCIi0relMk6ltG3ZzCLAbOCYbAYlIiJ9\n0x41f5nZG+5+WBbi2SNq/hIRSU8um78+m7QaAT5O0HHfY2Y2C/gZwXNabnP3WzrsPwu4EUiEr++4\n+/9momwREcm8VO7++j07+1RiBCPqf+vuG3tUsFkUeA84DVgDvALMdfclSccMcvf6cPkQ4M/uvn8n\n59KViohIGnJ2peLuX8x0oaGZwPKkucXmAWcBO5JKW0IJlQKbsxSLiIhkQCrNX78guFJpy2jtlt39\nyj0sexywKml9NXB0J+WfDdwElAOn72FZIiLSC1IZUV8EHADcS5BMzgPeBf7ew7JTaq9y94eAh8zs\nBOAuYFpnx1VWVu5YrqiooKKioofhiYj0H1VVVVRVVWW9nFT6VBYAx7t7a7ieD7zg7rtcVaRVsNkx\nQKW7zwrXrwUSHTvrO7znA2Bmx9H86lMREUlPtvpUUplQcijB1CxtysJtPfUqMMXMJplZAXABMD/5\nADPbz8wsXD4CYG+aHkZERNpLpfnrZuB1M3uWoPnrJKCypwW7e8zMrgCeJLil+HZ3X2Jml4X7bwU+\nC1xsZq0Ek1jO6Wm5IiKSPSkNfjSzcoJOdAcWuvu6bAeWDjV/iYikRxNKdkNJRUQkPbnsUxEREUmJ\nkoqIiGRMlx31Zja8uze6+9bMhyMiIn1Zd3d/vU73AxQnZzgWERHp49RRLyIyAOVsQsmw8GHAFIIp\nWwBw9+czHYyIiPRtqUwo+RXgSmA8sIjgqY8vAadkNzQREelrUrn76yqCaepXuPvJwOHA9qxGJSIi\nfVIqSaXJ3RsBzKzI3ZfSxUzBIiIysKXSp7Iq7FN5CHjazLYRPP1RRESknbTu/jKzCoIZi59w95Zs\nBZUu3f0lIpKenM39FT735F13rwnXBwMHuPuCTAezp5RURETSk8u5v35DMO18m/pwm4iISDspzf3l\n7omk5TjB809ERETaSSWpfGRmV5pZvpkVmNlVwIfZDkxERPqeVJLKV4FPAGuA1QSDH/8lm0GJiEjf\npLm/REQGoF6f+8vMrnb3W8zsF53sdne/MtPBiIhI39bd4Md3w7+vdthudD8lfsrMbBbwM4KO/9vc\n/ZYO+78A/FtYZi3wNXd/MxNli4hI5nXb/GVmUeA/3P3bGS84OPd7wGkE/TWvAHPdfUnSMccSjJHZ\nHiagSnc/ppNzqflLRCQNORmnEt4+/Akzy3jBBJNULnf3Fe7eCswDzupQ/kvu3jZ55QJg3yzEISIi\nGZLK3F9vAA+b2f1AQ7jN3f3BHpY9DliVtL4aOLqb478MPN7DMkVEJItSSSpFwFZ2fX5KT5NKyu1V\nZnYy8CWCW5s7VVlZuWO5oqKCioqKHoQmItK/VFVVUVVVlfVycnZLcTinWKW7zwrXrwUSnXTWH0qQ\nwGa5+/IuzqU+FRGRNPTHW4pfBaaY2SRgLXABMLdDDBMIEsqFXSUUERHZe6RyS/Frnezr8WWBu8fM\n7ArgSYJbim939yVmdlm4/1bgOmAY8OvwXoFWd5/Z07JFRCQ7Um7+MrMhBM1TtdkNKX1q/hIRSU/O\npr43s6PM7C3gTeBtM1tsZkdmOhAREen7UnlI11vA5e7+t3D9eOBX7n5oL8SXEl2piIikJ5cP6Yq1\nJRQAd38BiGU6EBER6ftSuVL5GVAM3BNuugBoAu4CcPfXsxlgKnSlIiKSnlw+o76K9nd7tZtQ0t1P\nznRQ6VJSERFJT86SSl+gpCIikp5eH/yYVPAw4GJgUtLxep6KiIjsIpW5vx4HXiK4pThBBp+nIiIi\n/UsqfSqvu/sRvRTPHlHzl4hIenLZUf+vQA3wCNDctt3dt2Y6mD2lpCIikp6c9akQ3D78Y+DfCZq/\nIGj++limgxERkb4tlSuVj4Cj3H1z74SUPl2piIikJ5cj6t8HGjNdsIiI9D+pNH81AG+Y2bPs7FPR\nLcUiIrKLVJLKQ+GrrX1JtxSLiEinUhpRb2YlwAR3X5r9kNKnPhURkfTk8nkqs4FFwBPh+uFmNj/T\ngYiISN+XSkd9JXA0sA3A3Reh24lFRKQTqSSVVnev7rAt0emRIiIyoKWSVN4xsy8AeWY2xcx+Afw9\nE4Wb2SwzW2pm75vZ1Z3sn25mL5lZk5l9OxNliohI9qSSVL4OHERwO/E9BFO2fKOnBZtZFPi/wCzg\nQGCumR3Q4bAtYfn/2dPyREQk+3Z7S7G71wPfDV+ZNBNY7u4rAMxsHnAWsCSp7E3AJjP7TIbLFhGR\nLEjlSiVbxgGrktZXh9tERKSPSmXwY7ZkdGBJZWXljuWKigoqKioyeXoRkT6tqqqKqqqqrJeTs8cJ\nm9kxQKW7zwrXrwUS7n5LJ8deD9S5+0+6OJcGP4qIpKHXp74P7/Jq4wTTs+xYz8DcX68CU8xsErAW\nuACY21U4PSxLRER6QXfNX6+Ff48juDvrXoIv9/OAd3pasLvHzOwK4EkgCtzu7kvM7LJw/61mNhZ4\nBRgMJMzsKuBAd6/rafkiIpJ5qTxPZQFwvLu3huv5wAvufnQvxJcSNX+JiKQnl89TGUpwpdCmLNwm\nIiLSTip3f90MvG5mVeH6SQTzgYmIiLST6tT35QSTSjqwwN3XZzuwdKj5S0QkPbls/mo7bhNQDUw1\nsxMzHYiIiPR9u23+MrNbCG73fReIJ+16PltBiYhI35TK3V/LgEPcvbnbA3NIzV8iIunJZfPXB0BB\npgsWEZH+J5W7vxqBN8zsrwTT30NmRtSLiEg/k0pSmR++kqmtSUREdpGzCSUzSX0qIiLp6fUJJZMK\nngr8iGD+r+Jws7v7xzIdjIiI9G2pdNT/DvgNEAMqgDuAP2QxJhER6aNSSSrF7v4MQVPZSnevBPR4\nXxER2UUqSaXJzKLAcjO7wszOBQZlOa498sgj0NSU6yhERAauVJLKN4AS4ErgSOBC4JJsBrWnZs+G\nn3T6bEgREekN/eruLwvvY+gHH0lEJKtyPaFkn2F68LCISM70r6RSaURGLct1FCnZuDFIgLW17bc/\n8USw3Qzuuis3sYmI7Kl+k1Ti4fzJU098Iyvnd4etWzNzrmXLYMzxj8E1QzhkRox77gEzx6KtfPqf\nF0ClQaVx8c33Y+bU1+/+nM/+rZEtWxMkEt0ft25TM0uWNwDQ3LKbg0VE0tRln4qZ/YJgOpbOGpQy\nMveXmc0CfgZEgdvc/ZZOjvk58GmgAfiiuy/q5BivqXEG/9Rg22T8Zx/2NLRdVFTAc8/Btm0wtMPD\nlO+9FwYNgn/6p2C9tRVaWuCRJxq58Oe/ID7zp3D/vXBpBTx8O5z15bTKjmw8jCM2/YTn7ziF4nD4\n6bJlcMpprawpewjOP7/rNz/3fUbMfJotxS+nXN6I+MH86cwqTjpqRLfHNTdDfj40NEBp6c7tDQ3B\nXXhFRfDhhzB9OuSlMiHQAPPOB9tpbopwxEFluQ4la4K+zt23STfFmqhprmFo0VDyInlsb9pOfWs9\nRXlFlBWUkR/NJ2LBb+CEJ9hYv5GN27dzf9W7bGxZQX1DghWbN1BWMJjyISMoLSomGoUDx4/j0x8/\nkOElw2iNt1JWWLbjPB3Vt9Tz3qYPGFY8hHFDyimIdj2Pbl1LHRvqNlBaUMqY0jF7Vjk5lq0+le6S\nSivwNnAfsLZtc/jX3f2OHhUc3Kb8HnAasAZ4BZjr7kuSjjkDuMLdzzCzo4H/dvdjOjmXb9zojP6V\nwfZ98Z+u6kloncc7fT7MPQsqHXf47/+G118PXm8X/xLihbDoUhi+HFrKYOhH8OXjuzzfJQddxpeP\nuogTf7/zmGFFI9jWtIX4dXEiFmFD3Qa++sgVPLTsgfaxxAvwaEu7bXPG/xt1iU18bNh+/Pzt73Va\n5qiici6e/nWGRsv50avfYeLgSSytfZU8LyZmjbscH2kayZDqE6ke9Tge7eJe7abBUFRDtH4fEtFG\nCmsPoGnU34N9sULIa8a2TqEwWsyo9XMpaZlMXcnbFEaKmGKzWFJwF5MbLiC2fjo+aB2bhzyFx/OJ\n5W2F/CbKC6YwKL+MSNNIPmx9gZroCibET2Zr9F1GxWdwxqTPMWeOUVgYXE0mElBeHiQ0d2hshC1b\n4A9/cF5a/i7R0m0YEcYW7Mf+5aOZNMk4/fTg2NJSiMVgwwYoKYEhQyDSzbW8e/ADYuPGYDkSgU2b\nnBcXbWF19TpaEy20JlpZt30LjfFaVtYvpTHWRKNXs6Hkr8RKVoElKKiZxuCWA3Cc5vx1FMZG4bFC\nEnl1lDCKZttKzJpxYiSijbTGWyhKjGGMHURhYhjR2BBi3kQ84RAvoKb4TYYVjOFAO5fJ+7VQYIPY\nb8xYXl3zGiu2rKO0YBBlhaUMLi6loaWRMYNHsLZ6E0Ut+3LwpDG8vfU1VlevZ0hxGVsatoJHqW+t\noSA/j/eq36TF62jyWkZEJxH1QmLeTGs8Tgt1xBPQmLea5vx1JCIttBatxRIFFDVNotDKGGYTKIiN\nZE3idVrztuKRFvAIzUX/CP4fyq+DSAJaBkGsCArqIBIDHDyKeR5OAlpKIZFP/pbDKC+YSkE+jBs6\nluqmbazZtplEpJmigijrG1eSGPVmcJ5EPoZh8WKizSOJtAzBYoPwSAvxgm3ES1fi28dBYQ0UV5PX\nPIqSxqmM4VASxInF4zRE1lNb8D7NxR9C0/AwvlYsXgTRVvJahxJNlFLSMolBNpKWRCMFlAEJ6j34\nvEUMY3hkPNGo8bHB02mNxWlM1FFWWEpZGRw4fAafOfJQRgzLIx6H2oZWttU18NZH69i/fAwjhwyi\npr6JBx/fzohx2zhwn0kcOq2MQYOMoUOhuDj4d7x5c/CjLz8fCgqCf5/19RCNBj8Ix4zp/aQyEjgP\nOJ/g4Vz3Ave7e3VGCjY7Frje3WeF69cAuPvNScf8BnjW3e8N15cCJ7n7hg7n8tWrnX1vM9g8Ff/F\ne2nFEo8HFd1RTQ185SvwtavqOPnpnb8mbzDn+hT/W1Rd8hw1zXXMnvcZEtclqG2ppaygLKVfb22a\nY8189t7zeGz5I+2233P2n7jg0HO6PVdrLE5+XicfrhvxRJyz7/kcjy5/aMe2Q4YdzfrGlUwo24/X\nNr2IYXz76GvZVFtNdfNmErECivMLWbj+ReYc9HkOHjOdRMIoLS7knjf+xEdbV7G2YQVFkUEkiNPQ\nWs/GlpW7lF1gRUQtn8ZELaPzJ7G1dS0xWijPm86EQdNp8M00tcQpiBawpP4FEi2F5DWVk8ivBRJE\n6scR27pv8KUzeDXEComM/IDE4KCssZGDMM9jQ2w5iWg91IyDVccFCXD7BBi8hug+b5LYOglvHAIf\nnkakbDOJaB37DZvCB2uq2XfEcDZsrad1xGIYuZS84gYi5JOINBAf+j5WVENZYiIJayGfEgZHRmMR\nZ2rpkRQXRRhaNJSKjx3PhSecQE1jA4++/hqLV79PhCijSkdQ0xR8URVFB/HRpnWMKB3KiEFDiHuM\nomgJxcUJlm9Yz0db/0FdYjONXk1JXikFkQJaaWR04UTWbNvEsxseYFB+KYloI/V5/2BU07GMKhjP\n9sR6zPNoTtSTb4U0+FaKbTgbG9YRL9zEaDuIYdEJNMSrKYkOIWIRBkWH0JpoYdrwAxhdOprSwmI+\n2PoRMW+iIFpISWERxfkF5OdFGF82kTGloynOL2J0STlxj/HB5n+wbmsN721ZSnXrRk7e73hGl42g\nMFpIa6KFw8ZPY+zIAhoaE+TlO4X5UaJRIx6H1liCppY4LbEYLa1xEgkYPyb4Au4u6bepqYH6hjh5\n0Sgbqmuorm+gNraVulg19bFaCqL5DCkcyvSR05i4zyBaWmDd5kaWrVvLG2vf5q21S8nLi1KUV8CI\n4hEcMHoaR+83jbHDy2hqctbXbGF7QwMeK2Rr/XZqGhtYvvUDVm1bx6iyYWxvqsHMGTt4FFHy2VCz\nlbV1q2mJxfmwZin50TxKIkOob62nqTnBOnuVhvyVWPMQwPDCasDIaywnXrSBRLQRSxTiTWXktQ4j\nXrYStxhgUFsOrcVgYGXrIJGHJyIQKw7anYq2Y/Hgx57ftL13k0q7g8z2BeYA3wKudvcedyGb2eeA\nT7n7V8L1C4Gj3f3rScc8Atzk7n8P158Jy3+tw7l84+XX8ct3boT3Z3HlZ49m+PDU4tiwAX79Gxgz\nGr72tWDb4sXw54eASAscfzNsmQYjlxK1CHHftR+ivHgiLV7PlqbNnLjvqbxX/RYjikcwbcQ0Zoyd\nkX7lDEDNsWbyInlEI6knwIQnWFe7juqmaupa6miOB+fYVL+ZWMxZuWUDB42bzORhE9mnbB8G5Q/a\nkYDdndZEK5vqN7GhfiPNsSY+WLuNEaWDmVpezvbmGrbV1fP26g8ZVFhCXjRCQ6yOhpoiikqbiEai\n7DNkNB8bVU5hXgGt8VYK8woZXDiE0YNGZaua9ljCE102+8jepyXeQlMsaB0ozismP5q/Y1/bd3bb\nv+V4Ik4i/F6qaa6lJd6KuzO0eDDujuO0xltJuFOcX7Tj38KgH/04ZxNKfpwgoXwS+AvwWvfvSFmq\no0k6fuhO3/flu5+lZhhQs5wDl4/h/JmTujxh5Q3B34svgjvvAsYsYsOGQ1mzJsrYsWFCATjhpuDv\nyKUMyi/hX4/7V2547kYADh59EAePPpgJQyZQnFfc7mrhFLpu9pLOFeYVpv2eiEUYN3gc4waPS/u9\nZkZBtKDd+48dv+txp07pH/8tlVD6loJoQZd9Oh1bJqKRKFGCH2MjSrr+NV21YgVVK97MXJBd6K75\n6wfAGcASYB7wpLu3Zqxgs2OAyqTmr2uBRHJnfdj8VeXu88L1Lpu/phy8nfc/NwSAv53qHN/Fd0E8\n3rHT2KEO3N+LAAAMTElEQVQyAtUTuWjLCmbNgi/c8AjnfHE1f265nENGH8KK6hVsu3rbjl/RqXY+\niojsrXIx+PHfgaHADOAm4HUzeyt8ZSLdvQpMMbNJZlYAXMCuDwObD1wMO5JQdceE0qY51rxj+Zln\nOi9w7Vq46KJwZehHwd/jwnldhq7kj3+Eb30L+Pxs/txyOQD3n3c/NdfWtGuWUUIREelcd81f3T0v\npccTobh7zMyuAJ4kuKX4dndfYmaXhftvdffHzewMM1sO1AOXdnW+5niYVNy44QaorNz1mCefhHvu\nAQ66D867AO59IOgzCcVHvcGG2n12rN906k1MGzmtpx9VRGTASHvuLwt+pp/fdkfW3sDMnOHvw5VT\ngg23bOa2X4zgkEOCAYuXXQb/+AfBrYkjl8Llh7R7/5kTv8AjK9s/Isav1wRiItJ/9Xrzl5mVmtm3\nzexXZna5mUXM7BzgHeALmQ6kx475r53Lec3887dXcvTR8Onv3co/vmQw/u9w7E93SSgAPz3zxnbr\nxXnFuxwjIiK7111H/YNADfAScDowHmgCrnT37MyFsofMzKlM2rD8U7D/k/DaP8PHb+v0PS99+SWO\nvf1Yvnz4l7lt9m08v/J5vvbY1/jtmb/liPIjKMor6pXYRURyIRcj6t9090PD5SiwDpjo7rsOvc4x\nM3O+vj+MWJ7S8XefczdfOHTvu9gSEekt2Uoq3XXUx9sW3D1uZmv2xoSyQ6yIfRs+w+qSx7o9rPX7\nreRFNBGViEg2dPfteqiZJU/MXpy07u4+OItxpS+vmajld7l7bOlY7j/vfiUUEZEs6rKj3t2j7l6W\n9MpLWt67EgpAJEYeBXyi/j932TX34Lk8c9EzHD+hf4yOFhHZW/Wfn+3R4ErlS+dM5sWndm5+8PwH\nmT1tdlpzSomIyJ7pPxMCRVvIswKKS4NBkPsP35/5c+ZzzgHnKKGIiPSS/pNUBm0mzwp23Aq8avsq\nzpx2Zo6DEhEZWPpPUgHyIvlMHDoRgEc//2iOoxERGXj6T58KkG8FHFF+BInrEpr0UUQkB/rZlUrw\n/AElFBGR3OhfSaWbcSoiIpJ9/Sqp5HfxpDQREekd/Sup6EpFRCSn+ldS0ZWKiEhO9aukUhBRUhER\nyaV+lVQwPa1RRCSX+lVSeWzlvFyHICIyoOUkqZjZcDN72syWmdlTZja0i+P+x8w2mNlbqZx334nN\nmQ1URETSkqsrlWuAp919KvDXcL0zvwNmpXrSsiGJDIQmIiJ7KldJZTZwR7h8B3B2Zwe5+9+Abame\nNOFKKiIiuZSrpDLG3TeEyxuAMZk46U2n3pSJ04iIyB7K2oSSZvY0MLaTXf+evOLubtbz27a+VP0l\nXr77ZV7mZSoqKqioqOjpKUVE+o2qqiqqqqqyXo659/5tuGa2FKhw9/VmVg486+7Tuzh2EvCIux/S\nzfk8F59DRKSvMjPcPeOz7+aq+Ws+cEm4fAnwUI7iEBGRDMpVUrkZ+KSZLQNOCdcxs33M7LG2g8zs\nHuDvwFQzW2Vml+YkWhERSUlOmr8yTc1fIiLp6W/NXyIi0g8pqYiISMYoqYiISMYoqYiISMYoqYiI\nSMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYoqYiISMYo\nqYiISMYoqYiISMYoqYiISMYoqYiISMbkJKmY2XAze9rMlpnZU2Y2tJNjxpvZs2b2jpm9bWZX5iJW\nERFJXa6uVK4Bnnb3qcBfw/WOWoFvuvtBwDHA/zGzA3oxxj6nqqoq1yHsNVQXO6kudlJdZF+uksps\n4I5w+Q7g7I4HuPt6d38jXK4DlgD79FqEfZD+h9lJdbGT6mIn1UX25SqpjHH3DeHyBmBMdweb2STg\ncGBBdsMSEZGeyMvWic3saWBsJ7v+PXnF3d3MvJvzlAIPAFeFVywiIrKXMvcuv8+zV6jZUqDC3deb\nWTnwrLtP7+S4fOBR4C/u/rNuztf7H0JEpI9zd8v0ObN2pbIb84FLgFvCvw91PMDMDLgdeLe7hALZ\nqRgREUlfrq5UhgP3AROAFcD57l5tZvsAv3X3z5jZ8cDzwJtAW5DXuvsTvR6wiIikJCdJRURE+qc+\nPaLezGaZ2VIze9/Mrs51PJnW1QDQ7gaPmtm1YX0sNbPTk7Z/3MzeCvf9dy4+TyaYWdTMFpnZI+H6\ngKwLMxtqZg+Y2RIze9fMjh7AdXFt+P/IW2b2RzMrHCh1YWb/Y2YbzOytpG0Z++xhXd4bbn/ZzCbu\nNih375MvIAosByYB+cAbwAG5jivDn3EscFi4XAq8BxwA/Afwb+H2q4Gbw+UDw3rID+tlOTuvRhcC\nM8Plx4FZuf58e1gn3wL+AMwP1wdkXRCM7/pSuJwHDBmIdRF+ng+BwnD9XoJ+2gFRF8AJBMMt3kra\nlrHPDlwO/CpcvgCYt7uY+vKVykxgubuvcPdWYB5wVo5jyijvfADoOLoePHoWcI+7t7r7CoJ/NEeH\nd9iVufvC8Lg76WTA6d7OzPYFzgBuA9puzhhwdWFmQ4AT3P1/ANw95u7bGYB1AdQQzL5RYmZ5QAmw\nlgFSF+7+N2Bbh82Z/OzJ5/oTcOruYurLSWUcsCppfXW4rV/qMAC0q8Gj+xDUQ5u2Oum4fQ19s67+\nC/gOkEjaNhDrYjKwycx+Z2avm9lvzWwQA7Au3H0r8BPgHwTJpNrdn2YA1kWSTH72Hd+z7h4Dtoc3\nWnWpLyeVAXOHQTgA9E8EA0Brk/d5cF3a7+vCzP4J2Ojui9h5ldLOQKkLguauIwiaJY4A6ukwf95A\nqQsz2w/4BkFzzj5AqZldmHzMQKmLzuTis/flpLIGGJ+0Pp722bZfCAeA/gm4y93bxvNsMLOx4f5y\nYGO4vWOd7EtQJ2vC5eTta7IZdxYcB8w2s4+Ae4BTzOwuBmZdrAZWu/sr4foDBElm/QCsiyOBv7v7\nlvCX9IPAsQzMumiTif8nVie9Z0J4rjxgSHh12KW+nFReBaaY2SQzKyDoRJqf45gyyqzLAaBtg0eh\n/eDR+cAcMysws8nAFGChu68HasI7hAy4iE4GnO7N3P277j7e3ScDc4D/dfeLGJh1sR5YZWZTw02n\nAe8AjzDA6gJYChxjZsXhZzgNeJeBWRdtMvH/xMOdnOtzBLPKdy/Xdy/08M6HTxPcEbWcYGBkzmPK\n8Oc7nqD/4A1gUfiaBQwHngGWAU8BQ5Pe892wPpYCn0ra/nHgrXDfz3P92XpYLyex8+6vAVkXwAzg\nFWAxwa/zIQO4Lv6NIKm+RdCpnD9Q6oLgqn0t0ELQ93FpJj87UEgwUP194GVg0u5i0uBHERHJmL7c\n/CUiInsZJRUREckYJRUREckYJRUREckYJRUREckYJRUREckYJRUZ8MzsxfDvRDObm+Fzf7ezskT6\nK41TEQmZWQXwbXc/M4335HkwPUhX+2vdvSwT8Yn0BbpSkQHPzOrCxZuBEyx4CNhVZhYxsx+b2UIz\nW2xm/xIeX2FmfzOzh4G3w20PmdmrFjxM7SvhtpuB4vB8dyWXZYEfhw9GetPMzk86d5WZ3W/BA7ju\nTorzZgseRrXYzH7cW/Ujko68XAcgshdou1y/GvjXtiuVMIlUu/tMMysEXjCzp8JjDwcOcveV4fql\n7r7NzIqBhWb2gLtfY2b/x90P76SscwmmWjkUGAW8YmbPh/sOI3ig0jrgRTP7BMG0Gme7+/QwtsGZ\nrQKRzNCVishOHafUPx242MwWEcx7NBzYP9y3MCmhAFxlZm8ALxHMBDtlN2UdD/zRAxuB54CjCJLO\nQndf60Hb9BvARKAaaDKz283sHKBxjz+lSBYpqYh07wp3Pzx87efuz4Tb69sOCPtiTgWOcffDCCb+\nLNrNeZ1dk1jbVUxz0rY4kO/ucYKnnT4A/BPwxJ58GJFsU1IR2akWSO5UfxK4PHyOBGY21cxKOnnf\nYGCbuzeZ2XTgmKR9rW3v7+BvwAVhv80o4ESC54R3+gCy8MmOQ939L8C3CJrORPY66lMR2XmFsBiI\nh81YvwN+TvBEwdfD50xsBM4Jj0++bfIJ4Ktm9i7BoxheStr3/4A3zew1D57/4gDu/mczOzYs04Hv\nuPtGMzuAXZ/U5wTJ7mEzKyJIPN/MyCcXyTDdUiwiIhmj5i8REckYJRUREckYJRUREckYJRUREckY\nJRUREckYJRUREckYJRUREckYJRUREcmY/w8l4SPJ8ZPJhwAAAABJRU5ErkJggg==\n", "text/html": [ "\n", "\n", "\n", "\n", "
\n", "" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "horizon = 10000\n", "data = S3(horizon)\n", "power_step = 1 # the limit law for power_step = 1/2 is not classical\n", "#######RM\n", "theta = 1\n", "path_RM = list()\n", "for i in range(horizon):\n", " gamma_k = 1./(i+1)**power_step # step size 1/sqrt(k)\n", " if theta > data[i]:\n", " s_k = 0.5 # 1-p quand p=1/2 pour la médiane\n", " else:\n", " s_k = -0.5\n", " theta = theta - gamma_k*s_k\n", " #print data[i], theta\n", " path_RM.append(theta)\n", "#######empirical quantile\n", "path_quant = []\n", "for i in range(horizon):\n", " path_quant.append(np.median(data[0:i]))\n", "########plot \n", "print 'horizon = {} \\n quantile = {}\\n RM = {}\\n theoretical median = 0'.format(horizon, path_quant[horizon-1], path_RM[horizon-1])\n", "#print len(iterations), len(path)\n", "plt.plot(range(horizon), path_quant, color = 'b', label = 'quantile')\n", "plt.plot(range(horizon), path_RM, color = 'g', label = 'RM')\n", "plt.xlabel(\"iterations\")\n", "plt.ylabel(\"RM and empirical quantile\")\n", "plt.plot(range(horizon), np.zeros(horizon),color = 'r')\n", "plt.legend(loc = 2)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Warm start for RM : start with $\\widehat q_{n,1/2}$ for $n=20$\n", "\n", "Open question : non-asymptotic convergence result for Robbins-Monro ?" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#Study of the Robbins-Monro algorithm for density $f_3$ (uniform over $[-0.5, 0.5]$) depending on the power in the step size ($=1/k^a$):\n", "\n", "$$x_{k+1} = x_k - \\eta_k(I(X_{k+1}\\leq x_k)-p) \\mbox{ for } \\eta_k=\\frac{1}{k^a} \\mbox{ and } a\\in[1/2,1]$$\n", "\n", "Asymptotic normality of the RM algorithm is known only when the step size is such that \n", "$$\\sum_k \\eta_k = +\\infty \\mbox{ and } \\sum_k \\eta_k^2<\\infty$$" ] }, { "cell_type": "code", "execution_count": 21, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeYAAAE4CAYAAACOp1CQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n", "text/html": [ "\n", "\n", "\n", "\n", "
\n", "" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import IPython.html.widgets as widgets\n", "from IPython.html.widgets import interact, interactive, fixed\n", "horizon = 5000\n", "theta_0=-2\n", "nb_paths = 1000\n", "\n", "@interact(a=(0.5, 1))\n", "def generate(a):\n", " list_theta_limit = list()\n", " for i in range(nb_paths):\n", " path = RM(theta_0, horizon, 3, power_step=a)\n", " list_theta_limit.append(path[horizon-1])\n", " ####Gaussian limit\n", " Fx = 0\n", " var = 1/float(4)\n", " x = np.linspace(min(list_theta_limit), max(list_theta_limit), 1000)\n", " y = norm.pdf(x, loc = Fx, scale = np.sqrt(var/float(horizon)))\n", " ####Plot\n", " f, ax = plt.subplots(1, 1, figsize = figsize)\n", " ax.hist(list_theta_limit, normed=True, bins=50)\n", " ax.plot(x,y)\n", " plt.show() " ] }, { "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": "2.7.10" } }, "nbformat": 4, "nbformat_minor": 0 }