{ "metadata": { "name": "" }, "nbformat": 3, "nbformat_minor": 0, "worksheets": [ { "cells": [ { "cell_type": "heading", "level": 1, "metadata": {}, "source": [ "Distributed Bayesian Logistic Regression" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Example for the paper _K. Dedecius and V. Se\u010dk\u00e1rov\u00e1: Distributed Modelling of Big Dynamic Data with Generalized Linear Models_**\n", "\n", "**This example demonstrates the linear regression with five cooperating processing nodes without a fusion center.**" ] }, { "cell_type": "heading", "level": 2, "metadata": {}, "source": [ "A bit on the logistic regression" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Observed random variable $y_k$ (0's and 1's) obeys Bernoulli distribution, that is, it's 1 with probability $p_k$ and 0 with $(1-p_k)$. The model reads\n", "\n", "$$\n", "y_k \\sim Bernoulli(p_k) \\qquad \\text{where} \\qquad logit(p_k) = \\log \\frac{p_k}{1-p_k} = \\theta' x_k,\n", "$$\n", "\n", "where\n", "\n", "$$\n", "\\mathbb{E}[y_k|x_k, \\theta] = p_k = logit^{-1}(\\theta' x_k) = \\frac{1}{1+e^{-\\theta' x_k}}.\n", "$$\n", "\n", "$x_k$ is the observed regressor. The goal is to estimate the vector of regression coefficients $\\theta$.\n", "\n", "The likelihood has the form\n", "\n", "$$\n", "L(\\theta, x_{1:k}, y_{1:k}) = \\prod_{i=1}^{k} p^{y_i} (1-p_i)^{1-y_i} = \n", "\\prod_{i=1}^{k} \\left(\\frac{1}{1+e^{-\\theta'x_i}}\\right)^{y_i}\n", " \\left(1-\\frac{1}{1+e^{-\\theta'x_i}}\\right)^{1-y_i}\n", "$$\n", "\n", "And the log-likelihood:\n", "\n", "$$\n", "l(\\theta, x_{1:k}, y_{1:k}) = \\sum_{i=1}^{k} [y_i \\log(p_i) + (1-y_i) \\log(1-p_i)].\n", "$$" ] }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Prior" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We have a normal prior for $\\theta$ (let us consider it scalar for this exposition). Hence\n", "\n", "$$\n", "\\theta \\sim \\mathcal{N}(\\mu_0, \\Sigma_0)\n", "$$\n", "\n", "with a pdf\n", "\n", "$$\n", "\\pi(\\theta) \\propto \\exp \\left\\{ \\frac{(\\theta - \\mu_0)^2}{2 \\Sigma_0} \\right\\}\n", "$$" ] }, { "cell_type": "heading", "level": 3, "metadata": {}, "source": [ "Posterior" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The posterior is not analytically reachable, but we can do [MAP estimation](http://en.wikipedia.org/wiki/Maximum_a_posteriori_estimation). The pdf is found via the Laplace approximation in the following steps:\n", "\n", "1. Search the mode of the posterior pdf. There will be the mode (= mean) of the approximating normal pdf.\n", "2. Set the variance of the approximating pdf. It is (perhaps generally) known, that it is the negative inverse second derivative of the log-pdf." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let us remind, that\n", "\n", "$$\\pi(\\theta|x_{1:k}, y_{1:k}) \\propto L(\\theta, x_{1:k}, y_{1:k}) \\pi(\\theta)$$\n", "\n", "We can find the mode using the [Newton's iterative method](http://en.wikipedia.org/wiki/Newton%27s_method_in_optimization), starting from some point $\\mu_k$ and iteratively improve by\n", "\n", "$$\n", "\\mu_k \\leftarrow \\mu_k - \\frac{\\partial_{\\theta}\\pi(\\theta|x_{1:k},y_{1:k})}{\\partial_{\\theta}^{2}\\pi(\\theta|x_{1:k},y_{1:k})} \n", "$$\n", "\n", "Note the ratio of derivatives only expresses approximation **stepsize**. Choosing the starting point to coincide with the **prior** mean, things simplify a bit, since we derive the stepsize only from the likelihood (recall $(\\theta - \\mu)$ in the exponent of the normal distribution)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This is clearly equivalent to (indices omitted)\n", "\n", "$$\n", "\\frac{d \\log (L(\\theta, x, y))}{d\\theta} = \\left(y - \\frac{1}{1-e^{-\\theta x}}\\right) x\n", "= (y - \\hat{y})x\n", "$$" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Second derivative of the log-likelihood:" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "$$\n", "\\frac\n", "{d^2 \\log(L(\\theta, x, y))}\n", "{d\\theta^2}\n", "= -x^2 \\hat{y}(1-\\hat{y})\n", "$$" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The Laplace approximation consists in (i) setting the *mode* of the approximating *normal* pdf to the mode of the approximated function and (ii) setting variance to the *negative inverse of the second derivative of the approximated (logarithmed) function*.\n", "\n", "This time, we need to involve the prior in our computations. But here the second derivative is really trivial. Thus, the variance of the posterior-approximating normal pdf is $[1/\\Sigma + x^2\\hat{y}(1-\\hat{y})]^{-1}$. Hence,\n", "\n", "$$\n", "\\pi(\\theta|x_{1:k}, y_{1:k}) \\approx \\mathcal{N}\\left(\\mu_k, \\left[\\frac{1}{\\Sigma}_k + x_k^2\\hat{y}_k(1-\\hat{y}_k)\\right]^{-1}\\right)\n", "$$" ] }, { "cell_type": "heading", "level": 2, "metadata": {}, "source": [ "Example" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Below we have our paper's example. After initialization of the ipython notebook, we simulate or load real data." ] }, { "cell_type": "code", "collapsed": false, "input": [ "import numpy as np\n", "from scipy.special import expit, logit\n", "import pylab as plt\n", "plt.rcParams['axes.grid'] = True\n", "np.set_printoptions(precision=3)\n", "\n", "task = 'realdata' # 'realdata' or 'simulation'\n", "coop_mode = 'cooperation' # 'cooperation' or 'nocoop'\n", "save_figs = False # Save figures?\n", "\n", "\n", "if task == 'simulation':\n", " ndat = 5000\n", " theta = np.array([-1.6, .03])\n", " np.random.seed(1234)\n", " x = np.random.randint(low=18, high=60, size=ndat)\n", " X = np.c_[np.ones(ndat), x]\n", " p_x = expit(np.dot(X, theta))\n", " y = np.random.binomial(n=1, p=p_x)\n", "\n", "elif task == 'realdata':\n", " data = np.genfromtxt('svmguide1.csv', delimiter=',')\n", " np.random.seed(1234)\n", " np.random.shuffle(data)\n", " ndat = data.shape[0]\n", " x = data[:,1]\n", " X = np.c_[np.ones(ndat), data[:, 1:]]\n", " y = data[:,0]\n", "\n", "nparams = X.shape[1]" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 1 }, { "cell_type": "markdown", "metadata": {}, "source": [ "The GLMs in general are somewhat tricky. Let us look at the properties of the data. Often, some transformations and preprocessing would help a lot (but we don't do it, we are dealing with raw dynamic data)." ] }, { "cell_type": "code", "collapsed": false, "input": [ "plt.figure(figsize=(6,.8*nparams))\n", "for i in range(1,nparams):\n", " plt.subplot(2,nparams/2,i)\n", " plt.boxplot(X[:,i], vert=True)\n", "plt.savefig('{}-{}-boxplot.pdf'.format(task, coop_mode))" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEACAYAAABI5zaHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbNyJiHSIjTsRkQ6xcSci0iE27kREOvT/h+UJW9cDZ2UAAAAASUVORK5CYII=\n", "text": [ "" ] } ], "prompt_number": 2 }, { "cell_type": "heading", "level": 2, "metadata": {}, "source": [ "Setting the class" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For easier use and reuse, we prepare a class for the logistic regression. This allows uncluttered coding.\n", "Note: we use R instead of $\\Sigma$ for better readability." ] }, { "cell_type": "code", "collapsed": false, "input": [ "class LogReg():\n", " \"\"\"Logistic regression class\"\"\"\n", " \n", " def __init__(self, R, theta_hat):\n", " \"\"\"Constructor\"\"\"\n", " self.inv_R = np.linalg.inv(R) # Prior covariance R = \\Sigma\n", " self.theta_hat = theta_hat # Prior theta = \\mu\n", " self.forg_factor = 1. # Forgetting factor\n", " self.neigh_weights = [] # Neighbors' weights\n", " self.nupdates = 0 # No. of updates\n", " self.theta_hat_log = [] # Log of estimates\n", " self.var_log = [] # Log of variances\n", " self.merged_theta_hat_log = [] # (analog. merged estimates)\n", " self.merged_var_log = [] # dtto\n", " self.brier_log = [] # Log of Brier score\n", " self.brier_sum = 0. # Brier score - sum\n", " \n", " def update(self, y, X):\n", " \"\"\"Bayesian update of prior\"\"\"\n", " yhat = expit(np.dot(X, self.theta_hat))\n", " Dl = (y - yhat) * X\n", " D2l = -self.inv_R - yhat * (1. - yhat) * np.outer(X, X)\n", " theta_hat_post = self.theta_hat - np.dot(np.linalg.inv(D2l), Dl.T)\n", " self.inv_R = - D2l * self.forg_factor\n", " self.theta_hat = theta_hat_post # Estimate of the posterior mean\n", " self.nupdates += 1\n", " self.theta_hat_log.append(self.theta_hat)\n", " self.var_log.append(np.diag(np.linalg.inv(self.inv_R)))\n", " self.brier_sum += (y - yhat) ** 2\n", " self.brier_log.append(self.brier_score)\n", " \n", " def merge(self, theta_hats_list, inv_Rs_list):\n", " \"\"\"Fusion of posteriors\"\"\"\n", " inv_R_merged = np.zeros(self.inv_R.shape)\n", " theta_hat_merged = np.zeros(self.theta_hat.shape)\n", " for i in range(len(self.neigh_weights)):\n", " inv_R_merged += self.neigh_weights[i] * inv_Rs_list[i]\n", " for i in range(len(self.neigh_weights)):\n", " convex_comb = self.neigh_weights[i] \\\n", " * np.dot(inv_Rs_list[i], theta_hats_list[i])\n", " convex_comb = np.dot(np.linalg.inv(inv_R_merged), convex_comb.flatten())\n", " theta_hat_merged += convex_comb.flatten()\n", " self.theta_hat = theta_hat_merged\n", " self.inv_R = inv_R_merged\n", " self.merged_theta_hat_log.append(theta_hat_merged)\n", " self.merged_var_log.append(np.diag(np.linalg.inv(inv_R_merged)))\n", " \n", " @property\n", " def R(self):\n", " return np.linalg.inv(self.inv_R)\n", " \n", " @property\n", " def brier_score(self):\n", " assert self.nupdates > 0.\n", " return self.brier_sum / self.nupdates" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 3 }, { "cell_type": "markdown", "metadata": {}, "source": [ "The `update()` function does the Bayesian update, principally described above." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The `merge()` function does fusion. With weights $\\omega_{j}^{i} \\in [0,1]$ (summing to unity for fixed $i$), it results in\n", "\\begin{equation}\n", " \\tilde{\\mu}_{k,i} = \\tilde\\Sigma_{k,i} %\\left( \\sum_{j\\in\\mathcal{N}^{i}} \\omega_{j}^{i} \\Sigma_{j}^{-1} \\right)^{-1} \n", " \\left( \\sum_{j\\in\\mathcal{N}^{i}} \\omega_{j}^{i} \\Sigma_{k,j}^{-1}\\mu_{k,j}\\right)\n", " \\qquad\\text{and}\\qquad\n", " \\tilde\\Sigma_{k,i} = \\left( \\sum_{j\\in\\mathcal{N}^{i}} \\omega_{j}^{i} \\Sigma_{k,j}^{-1} \\right)^{-1}.\n", " \\notag\n", "\\end{equation}\n", "\n", "See the paper for the details." ] }, { "cell_type": "heading", "level": 2, "metadata": {}, "source": [ "Networked estimation" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The networked estimation runs with 5 nodes with uniform weights. The topology of the network depicts the following figure:" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We initialize with prior covariance\n", "$R = \\Sigma = N(0, diag(10))$ and set the neighbors." ] }, { "cell_type": "code", "collapsed": false, "input": [ "R_prior = np.eye(nparams) * 10.\n", "theta_hat_prior = np.zeros(nparams)\n", "\n", "nnodes = 5\n", "nodes = [LogReg(R_prior.copy(), theta_hat_prior.copy()) for i in range(nnodes)]\n", "\n", "# Assign weights\n", "weights = np.eye(nnodes)\n", "if coop_mode == 'cooperation':\n", " weights[0, (1,4)] = 1.\n", " weights[1, (0,2,3)] = 1.\n", " weights[2, (1,3,4)] = 1.\n", " weights[3, (1,2,4)] = 1.\n", " weights[4, (0,2,3)] = 1.\n", " weights /= weights.sum(axis=0)[:,np.newaxis]\n", "\n", "for i in range(nnodes):\n", " nodes[i].neigh_weights = weights[i] " ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 4 }, { "cell_type": "markdown", "metadata": {}, "source": [ "Below runs the estimation..." ] }, { "cell_type": "code", "collapsed": false, "input": [ "data_iter = iter(zip(y, X))\n", "have_data = True\n", "inv_Rs_list = []\n", "theta_hats_list = []\n", "common_data_len = 0\n", "while have_data == True:\n", " for node in nodes:\n", " try:\n", " yi, Xi = data_iter.next()\n", " except:\n", " have_data = False\n", " break\n", " node.update(yi, Xi)\n", " common_data_len += 1\n", " theta_hats_list = [node.theta_hat for node in nodes]\n", " inv_Rs_list = [node.inv_R for node in nodes]\n", " for node in nodes:\n", " node.merge(theta_hats_list, inv_Rs_list)" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 5 }, { "cell_type": "markdown", "metadata": {}, "source": [ "...preprocessing of logs..." ] }, { "cell_type": "code", "collapsed": false, "input": [ "theta_hats_log = np.array([node.merged_theta_hat_log[:common_data_len-1] for node in nodes])\n", "vars_log = np.array([node.merged_var_log[:common_data_len-1] for node in nodes])\n", "brier_log = np.column_stack([np.array(node.brier_log)[:common_data_len-1] for node in nodes])" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 6 }, { "cell_type": "markdown", "metadata": {}, "source": [ "...and plotting of results. First estimates for all nodes," ] }, { "cell_type": "code", "collapsed": false, "input": [ "plt.figure(figsize=(6,1.8*nparams))\n", "for i in range(nparams):\n", " plt.subplot(nparams,1,i+1)\n", " plt.plot(theta_hats_log[:,:,i].T)" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAIXCAYAAACy4ggxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n", "text": [ "" ] } ], "prompt_number": 7 }, { "cell_type": "markdown", "metadata": {}, "source": [ "Second variances (their logarithms)," ] }, { "cell_type": "code", "collapsed": false, "input": [ "plt.figure(figsize=(6,1.8*nparams))\n", "for i in range(nparams):\n", " plt.subplot(nparams,1,i+1)\n", " plt.plot(np.log(vars_log[:,:,i].T))" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAIXCAYAAAB+ezLaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text": [ "" ] } ], "prompt_number": 8 }, { "cell_type": "heading", "level": 2, "metadata": {}, "source": [ "Publication figures" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The publication presents evolution of parameter estimates, their variances and the Brier score for a single node only. The figures come the following code:" ] }, { "cell_type": "code", "collapsed": false, "input": [ "plt.figure(figsize=(6,1.8*nparams))\n", "for i in range(nparams):\n", " plt.subplot(nparams,1,i+1)\n", " plt.plot(theta_hats_log[1,:,i].T)\n", " plt.hold(True)\n", " plt.plot(theta_hats_log[1,:,i] + 3. * np.sqrt(vars_log[1,:,i]), 'g')\n", " plt.plot(theta_hats_log[1,:,i] - 3. * np.sqrt(vars_log[1,:,i]), 'g')\n", " plt.xlim(xmax=common_data_len)\n", " plt.ylabel(r'$\\widehat\\Theta_{0}$'.format(i))\n", "plt.savefig('{0}-{1}-est.pdf'.format(task, coop_mode)) if save_figs else None" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAIXCAYAAABKAvPGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text": [ "" ] } ], "prompt_number": 9 }, { "cell_type": "code", "collapsed": false, "input": [ "plt.figure(figsize=(6,2.9))\n", "for i in range(nnodes):\n", " plt.plot(np.log(brier_log[:,i]), label=r'Node{0}'.format(i+1))\n", " plt.hold(True)\n", "plt.xlabel('# of data per node')\n", "plt.ylabel('Log Brier score')\n", "plt.legend(borderaxespad=0, ncol=3, bbox_to_anchor=(0, 0, 1, 1))\n", "plt.xlim(xmax=common_data_len)\n", "#plt.ylim(ymin=-1.6, ymax=-0.6)\n", "plt.savefig('{0}-{1}-brier.pdf'.format(task, coop_mode)) if save_figs else None\n", "print('Final Brier score: {}'.format(brier_log[-1, :]))\n" ], "language": "python", "metadata": {}, "outputs": [ { "output_type": "stream", "stream": "stdout", "text": [ "Final Brier score: [ 0.033 0.041 0.037 0.039 0.035]\n" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAADSCAYAAABdPydEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBAKBoFcIwyEQCASCXiEMh0AgEAh6xf8B2z3tNMHAEe0AAAAASUVORK5CYII=\n", "text": [ "" ] } ], "prompt_number": 10 }, { "cell_type": "markdown", "metadata": {}, "source": [ "With simulated data, we also know the estimates' MSEs:" ] }, { "cell_type": "code", "collapsed": false, "input": [ "if task == 'simulation':\n", " print (((theta_hats_log - theta) ** 2)/float(common_data_len - 1)).sum(axis=1)" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 11 } ], "metadata": {} } ] }