{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Bayesian regression with linear basis function models\n", "\n", "This article is an introduction to Bayesian regression with linear basis function models. After a short overview of the relevant mathematical results and their intuition, Bayesian linear regression is implemented from scratch with [NumPy](http://www.numpy.org/) followed by an example how [scikit-learn](https://scikit-learn.org/stable/) can be used to obtain equivalent results. It is assumed that you already have a basic understanding probability distributions and [Bayes' theorem](https://en.wikipedia.org/wiki/Bayes%27_theorem). For a detailed mathematical coverage I recommend reading chapter 3 of [Pattern Recognition and Machine Learning](https://www.microsoft.com/en-us/research/uploads/prod/2006/01/Bishop-Pattern-Recognition-and-Machine-Learning-2006.pdf) (PRML) but this is not necessary for following this article. \n", "\n", "## Linear basis function models\n", "\n", "Linear regression models share the property of being linear in their parameters but not necessarily in their input variables. Using non-linear basis functions of input variables, linear models are able model arbitrary non-linearities from input variables to targets. Polynomial regression is such an example and will be demonstrated later. A linear regression model $y(\\mathbf{x}, \\mathbf{w})$ can therefore be defined more generally as\n", "\n", "$$\n", "y(\\mathbf{x}, \\mathbf{w}) = w_0 + \\sum_{j=1}^{M-1}{w_j \\phi_j(\\mathbf{x})} = \\sum_{j=0}^{M-1}{w_j \\phi_j(\\mathbf{x})} = \\mathbf{w}^T \\boldsymbol\\phi(\\mathbf{x}) \\tag{1}\n", "$$\n", "\n", "where $\\phi_j$ are basis functions and $M$ is the total number of parameters $w_j$ including the bias term $w_0$. Here, we use the convention $\\phi_0(\\mathbf{x}) = 1$. The simplest form of linear regression models are also linear functions of their input variables i.e. the set of basis functions in this case is the identity $\\boldsymbol\\phi(\\mathbf{x}) = \\mathbf{x}$. The target variable $t$ of an observation $\\mathbf{x}$ is given by a deterministic function $y(\\mathbf{x}, \\mathbf{w})$ plus additive random noise $\\epsilon$. \n", "\n", "$$\n", "t = y(\\mathbf{x}, \\mathbf{w}) + \\epsilon \\tag{2}\n", "$$\n", "\n", "We make the assumption that the noise is normally distributed i.e. follows a Gaussian distribution with zero mean and precision (= inverse variance) $\\beta$. The corresponding probabilistic model i.e. the conditional distribution of $t$ given $\\mathbf{x}$ can therefore be written as\n", "\n", "$$\n", "p(t \\lvert \\mathbf{x}, \\mathbf{w}, \\beta) = \n", "\\mathcal{N}(t \\lvert y(\\mathbf{x}, \\mathbf{w}), \\beta^{-1}) =\n", "\\sqrt{\\beta \\over {2 \\pi}} \\exp\\left(-{\\beta \\over 2} (t - y(\\mathbf{x}, \\mathbf{w}))^2 \\right) \\tag{3}\n", "$$\n", "\n", "where the mean of this distribution is the regression function $y(\\mathbf{x}, \\mathbf{w})$. \n", "\n", "## Likelihood function\n", "\n", "For fitting the model and for inference of model parameters we use a training set of $N$ independent and identically distributed (i.i.d.) observations $\\mathbf{x}_1,\\ldots,\\mathbf{x}_N$ and their corresponding targets $t_1,\\ldots,t_N$. After combining column vectors $\\mathbf{x}_i$ into matrix $\\mathbf{X}$, where $\\mathbf{X}_{i,:} = \\mathbf{x}_i^T$, and scalar targets $t_i$ into column vector $\\mathbf{t}$ the joint conditional probability of targets $\\mathbf{t}$ given $\\mathbf{X}$ can be formulated as\n", "\n", "$$\n", "p(\\mathbf{t} \\lvert \\mathbf{X}, \\mathbf{w}, \\beta) = \n", "\\prod_{i=1}^{N}{\\mathcal{N}(t_i \\lvert \\mathbf{w}^T \\boldsymbol\\phi(\\mathbf{x}_i), \\beta^{-1})} \\tag{4}\n", "$$\n", "\n", "This is a function of parameters $\\mathbf{w}$ and $\\beta$ and is called the *likelihood function*. For better readability, it will be written as $p(\\mathbf{t} \\lvert \\mathbf{w}, \\beta)$ instead of $p(\\mathbf{t} \\lvert \\mathbf{X}, \\mathbf{w}, \\beta)$ from now on. The log of the likelihood function can be written as \n", "\n", "$$\n", "\\log p(\\mathbf{t} \\lvert \\mathbf{w}, \\beta) = \n", "{N \\over 2} \\log \\beta - \n", "{N \\over 2} \\log {2 \\pi} - \n", "\\beta E_D(\\mathbf{w}) \\tag{5}\n", "$$\n", "\n", "where $E_D(\\mathbf{w})$ is the sum-of-squares error function coming from the exponent of the likelihood function.\n", "\n", "$$\n", "E_D(\\mathbf{w}) = \n", "{1 \\over 2} \\sum_{i=1}^{N}(t_i - \\mathbf{w}^T \\boldsymbol\\phi(\\mathbf{x}_i))^2 = \n", "{1 \\over 2} \\lVert \\mathbf{t} - \\boldsymbol\\Phi \\mathbf{w} \\rVert^2 \\tag{6}\n", "$$\n", "\n", "Matrix $\\boldsymbol\\Phi$ is called the *design matrix* and is defined as\n", "\n", "$$\n", "\\boldsymbol\\Phi = \n", "\\begin{pmatrix}\n", "\\phi_0(\\mathbf{x}_1) & \\phi_1(\\mathbf{x}_1) & \\cdots & \\phi_{M-1}(\\mathbf{x}_1) \\\\ \n", "\\phi_0(\\mathbf{x}_2) & \\phi_1(\\mathbf{x}_2) & \\cdots & \\phi_{M-1}(\\mathbf{x}_2) \\\\\n", "\\vdots & \\vdots & \\ddots & \\vdots \\\\\n", "\\phi_0(\\mathbf{x}_N) & \\phi_1(\\mathbf{x}_N) & \\cdots & \\phi_{M-1}(\\mathbf{x}_N)\n", "\\end{pmatrix} \\tag{7}\n", "$$\n", "\n", "## Maximum likelihood\n", "\n", "Maximizing the log likelihood (= minimizing the sum-of-squares error function) w.r.t. $\\mathbf{w}$ gives the maximum likelihood estimate of parameters $\\mathbf{w}$. Maximum likelihood estimation can lead to severe over-fitting if complex models (e.g. polynomial regression models of high order) are fit to datasets of limited size. A common approach to prevent over-fitting is to add a regularization term to the error function. As we will see shortly, this regularization term arises naturally when following a Bayesian approach (more precisely, when defining a prior distribution over parameters $\\mathbf{w}$). \n", "\n", "## Bayesian approach\n", "\n", "### Prior and posterior distribution\n", "\n", "For a Bayesian treatment of linear regression we need a prior probability distribution over model parameters $\\mathbf{w}$. For reasons of simplicity, we will use an isotropic Gaussian distribution over parameters $\\mathbf{w}$ with zero mean:\n", "\n", "$$\n", "p(\\mathbf{w} \\lvert \\alpha) = \\mathcal{N}(\\mathbf{w} \\lvert \\mathbf{0}, \\alpha^{-1}\\mathbf{I}) \\tag{8}\n", "$$\n", "\n", "An isotropic Gaussian distribution has a diagonal covariance matrix where all diagonal elements have the same variance $\\alpha^{-1}$ ($\\alpha$ is the precision of the prior). A zero mean favors small(er) values of parameters $w_j$ a priori. The prior is [conjugate](https://en.wikipedia.org/wiki/Conjugate_prior) to the likelihood $p(\\mathbf{t} \\lvert \\mathbf{w}, \\beta)$ meaning that the posterior distribution has the same functional form as the prior i.e. is also a Gaussian. In this special case, the posterior has an analytical solution with the following sufficient statistics \n", "\n", "$$\n", "\\begin{align*}\n", "\\mathbf{m}_N &= \\beta \\mathbf{S}_N \\boldsymbol\\Phi^T \\mathbf{t} \\tag{9} \\\\\n", "\\mathbf{S}_N^{-1} &= \\alpha\\mathbf{I} + \\beta \\boldsymbol\\Phi^T \\boldsymbol\\Phi \\tag{10}\n", "\\end{align*}\n", "$$\n", "\n", "$(9)$ is the mean vector of the posterior and $(10)$ the inverse covariance matrix (= precision matrix). Hence, the posterior distribution can be written as\n", "\n", "$$\n", "p(\\mathbf{w} \\lvert \\mathbf{t}, \\alpha, \\beta) = \\mathcal{N}(\\mathbf{w} \\lvert \\mathbf{m}_N, \\mathbf{S}_N) \\tag{11}\n", "$$\n", "\n", "For the moment, we assume that the values of $\\alpha$ and $\\beta$ are known. Since the posterior is proportional to the product of likelihood and prior, the log of the posterior distribution is proportional to the sum of the log likelihood and the log of the prior\n", "\n", "$$\n", "\\log p(\\mathbf{w} \\lvert \\mathbf{t}, \\alpha, \\beta) = \n", "-\\beta E_D(\\mathbf{w}) - \\alpha E_W(\\mathbf{w}) + \\mathrm{const.} \\tag{12}\n", "$$\n", "\n", "where $E_D(\\mathbf{w})$ is defined by $(6)$ and \n", "\n", "$$\n", "E_W(\\mathbf{w}) = {1 \\over 2} \\mathbf{w}^T \\mathbf{w} \\tag{13}\n", "$$\n", "\n", "Maximizing the log posterior w.r.t. $\\mathbf{w}$ gives the [maximum-a-posteriori](https://en.wikipedia.org/wiki/Maximum_a_posteriori_estimation) (MAP) estimate of $\\mathbf{w}$. Maximizing the log posterior is equivalent to minimizing the sum-of-squares error function $E_D$ plus a quadratic regularization term $E_W$. This particular form regularization is known as *L2 regularization* or *weight decay* as it limits the magnitude of weights $w_j$. The contribution of the regularization term is determined by the ratio $\\alpha / \\beta$.\n", "\n", "### Posterior predictive distribution\n", "\n", "For making a prediction $t$ at a new location $\\mathbf{x}$ we use the posterior predictive distribution which is defined as\n", "\n", "$$\n", "p(t \\lvert \\mathbf{x}, \\mathbf{t}, \\alpha, \\beta) = \n", "\\int{p(t \\lvert \\mathbf{x}, \\mathbf{w}, \\beta) p(\\mathbf{w} \\lvert \\mathbf{t}, \\alpha, \\beta) d\\mathbf{w}} \\tag{14}\n", "$$\n", "\n", "The posterior predictive distribution includes uncertainty about parameters $\\mathbf{w}$ into predictions by weighting the conditional distribution $p(t \\lvert \\mathbf{x}, \\mathbf{w}, \\beta)$ with the posterior probability of weights $p(\\mathbf{w} \\lvert \\mathbf{t}, \\alpha, \\beta)$ over the entire weight parameter space. By using the predictive distribution we're not only getting the expected value of $t$ at a new location $\\mathbf{x}$ but also the uncertainty for that prediction. In our special case, the posterior predictive distribution is a Gaussian distribution\n", "\n", "$$\n", "p(t \\lvert \\mathbf{x}, \\mathbf{t}, \\alpha, \\beta) = \n", "\\mathcal{N}(t \\lvert \\mathbf{m}_N^T \\boldsymbol\\phi(\\mathbf{x}), \\sigma_N^2(\\mathbf{x})) \\tag{15}\n", "$$\n", "\n", "where mean $\\mathbf{m}_N^T \\boldsymbol\\phi(\\mathbf{x})$ is the regression function after $N$ observations and $\\sigma_N^2(\\mathbf{x})$ is the corresponding predictive variance\n", "\n", "$$\n", "\\sigma_N^2(\\mathbf{x}) = {1 \\over \\beta} + \\boldsymbol\\phi(\\mathbf{x})^T \\mathbf{S}_N \\boldsymbol\\phi(\\mathbf{x}) \\tag{16}\n", "$$\n", "\n", "The first term in $(16)$ represents the inherent noise in the data and the second term covers the uncertainty about parameters $\\mathbf{w}$. So far, we have assumed that the values of $\\alpha$ and $\\beta$ are known. In a fully Bayesian treatment, however, we should define priors over $\\alpha$ and $\\beta$ and use the corresponding posteriors to additionally include uncertainties about $\\alpha$ and $\\beta$ into predictions. Unfortunately, complete integration over all three parameters $\\mathbf{w}$, $\\alpha$ and $\\beta$ is analytically intractable and we have to use another approach.\n", "\n", "### Evidence function\n", "\n", "Estimates for $\\alpha$ and $\\beta$ can alternatively be obtained by first integrating the product of likelihood and prior over parameters $\\mathbf{w}$\n", "\n", "$$\n", "p(\\mathbf{t} \\lvert \\alpha, \\beta) =\n", "\\int{p(\\mathbf{t} \\lvert \\mathbf{w}, \\beta) p(\\mathbf{w} \\lvert \\alpha) d\\mathbf{w}} \\tag{17}\n", "$$\n", "\n", "and then maximizing the resulting *marginal likelihood* or *evidence function* w.r.t. $\\alpha$ and $\\beta$. This approach is known as [empirical Bayes](https://en.wikipedia.org/wiki/Empirical_Bayes_method). It can be shown that this is a good approximation for a fully Bayesian treatment if the posterior for $\\alpha$ and $\\beta$ is sharply peaked around the most probable value and the prior is relatively flat which is often a reasonable assumption. Integrating over model parameters or using a good approximation for it allows us to estimate values for $\\alpha$ and $\\beta$, and hence the regularization strength $\\alpha / \\beta$, from training data alone i.e. without using a validation set. \n", "\n", "The log of the marginal likelihood is given by\n", "\n", "$$\n", "\\log p(\\mathbf{t} \\lvert \\alpha, \\beta) = {M \\over 2} \\log \\alpha + {N \\over 2} \\log \\beta -\n", "E(\\mathbf{m}_N) - {1 \\over 2} \\log \\lvert \\mathbf{S}_N^{-1}\\rvert - {N \\over 2} \\log {2 \\pi} \\tag{18}\n", "$$\n", "\n", "where\n", "\n", "$$\n", "E(\\mathbf{m}_N) = {\\beta \\over 2} \\lVert \\mathbf{t} - \\boldsymbol\\Phi \\mathbf{m}_N \\rVert^2 +\n", "{\\alpha \\over 2} \\mathbf{m}_N^T \\mathbf{m}_N \\tag{19}\n", "$$\n", "\n", "For completeness, the relationship between evidence, likelihood, prior, posterior is of course given by Bayes' theorem\n", "\n", "$$\n", "p(\\mathbf{w} \\lvert \\mathbf{t}, \\alpha, \\beta) = \n", "{p(\\mathbf{t} \\lvert \\mathbf{w}, \\beta) p(\\mathbf{w} \\lvert \\alpha) \\over p(\\mathbf{t} \\lvert \\alpha, \\beta)} \\tag{20}\n", "$$\n", "\n", "#### Maximization\n", "\n", "Maximization of the log marginal likelihood w.r.t. $\\alpha$ and $\\beta$ gives the following implicit solutions.\n", "\n", "$$\n", "\\alpha = {\\gamma \\over \\mathbf{m}_N^T \\mathbf{m}_N} \\tag{21}\n", "$$\n", "\n", "and \n", "\n", "$$\n", "{1 \\over \\beta} = {1 \\over N - \\gamma} \\sum_{i=1}^{N}(t_i - \\mathbf{m}_N^T \\boldsymbol\\phi(\\mathbf{x}_i))^2 \\tag{22}\n", "$$\n", "\n", "where \n", "\n", "$$\n", "\\gamma = \\sum_{i=0}^{M-1} {\\lambda_i \\over \\alpha + \\lambda_i} \\tag{23}\n", "$$\n", "\n", "\n", "and $\\lambda_i$ are the [eigenvalues](https://en.wikipedia.org/wiki/Eigenvalues_and_eigenvectors) of $\\beta \\boldsymbol\\Phi^T \\boldsymbol\\Phi$. The solutions are implicit because $\\alpha$ and $\\gamma$ as well as $\\beta$ and $\\gamma$ depend on each other. Solutions for $\\alpha$ and $\\beta$ can therefore be obtained by starting with initial values for these parameters and then iterating over the above equations until convergence.\n", "\n", "#### Evaluation\n", "\n", "Integration over model parameters also makes models of different complexity directly comparable by evaluating their evidence function on training data alone without needing a validation set. Further below we'll see an example how polynomial models of different complexity (i.e. different polynomial degree) can be compared directly by evaluating their evidence function alone. The highest evidence is usually obtained for models of intermediate complexity i.e. for models whose complexity is just high enough for explaining the data sufficiently well. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Implementation\n", "\n", "### Posterior and posterior predictive distribution\n", "\n", "We start with the implementation of the posterior and posterior predictive distributions. Function `posterior` computes the mean and covariance matrix of the posterior distribution and function `posterior_predictive` computes the mean and the variances of the posterior predictive distribution. Here, readability of code and similarity to the mathematical definitions has higher priority than optimizations." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "\n", "\n", "def posterior(Phi, t, alpha, beta, return_inverse=False):\n", " \"\"\"Computes mean and covariance matrix of the posterior distribution.\"\"\"\n", " S_N_inv = alpha * np.eye(Phi.shape[1]) + beta * Phi.T.dot(Phi)\n", " S_N = np.linalg.inv(S_N_inv)\n", " m_N = beta * S_N.dot(Phi.T).dot(t)\n", "\n", " if return_inverse:\n", " return m_N, S_N, S_N_inv\n", " else:\n", " return m_N, S_N\n", "\n", "\n", "def posterior_predictive(Phi_test, m_N, S_N, beta):\n", " \"\"\"Computes mean and variances of the posterior predictive distribution.\"\"\"\n", " y = Phi_test.dot(m_N)\n", " # Only compute variances (diagonal elements of covariance matrix)\n", " y_var = 1 / beta + np.sum(Phi_test.dot(S_N) * Phi_test, axis=1)\n", " \n", " return y, y_var" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Example datasets\n", "\n", "The datasets used in the following examples are based on $N$ scalar observations $x_{i = 1,\\ldots,N}$ which are combined into a $N \\times 1$ matrix $\\mathbf{X}$. Target values $\\mathbf{t}$ are generated from $\\mathbf{X}$ with functions `f` and `g` which also generate random noise whose variance can be specified with the `noise_variance` parameter. We will use `f` for generating noisy samples from a straight line and `g` for generating noisy samples from a sinusoidal function." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "f_w0 = -0.3\n", "f_w1 = 0.5\n", "\n", "\n", "def f(X, noise_variance):\n", " '''Linear function plus noise'''\n", " return f_w0 + f_w1 * X + noise(X.shape, noise_variance)\n", "\n", "\n", "def g(X, noise_variance):\n", " '''Sinusoidial function plus noise'''\n", " return 0.5 + np.sin(2 * np.pi * X) + noise(X.shape, noise_variance)\n", "\n", "\n", "def noise(size, variance):\n", " return np.random.normal(scale=np.sqrt(variance), size=size)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Basis functions\n", "\n", "For straight line fitting, a model that is linear in its input variable $x$ is sufficient. Hence, we don't need to transform $x$ with a basis function which is equivalent to using an `identity_basis_function`. For fitting a linear model to a sinusoidal dataset we transform input $x$ with `gaussian_basis_function` and later with `polynomial_basis_function`. These non-linear basis functions are necessary to model the non-linear relationship between input $x$ and target $t$. The design matrix $\\boldsymbol\\Phi$ can be computed from observations $\\mathbf{X}$ and a parametric basis function with function `expand`. This function also prepends a column vector $\\mathbf{1}$ according to $\\phi_0(x) = 1$." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "def identity_basis_function(x):\n", " return x\n", "\n", "\n", "def gaussian_basis_function(x, mu, sigma=0.1):\n", " return np.exp(-0.5 * (x - mu) ** 2 / sigma ** 2)\n", "\n", "\n", "def polynomial_basis_function(x, power):\n", " return x ** power\n", "\n", "\n", "def expand(x, bf, bf_args=None):\n", " if bf_args is None:\n", " return np.concatenate([np.ones(x.shape), bf(x)], axis=1)\n", " else:\n", " return np.concatenate([np.ones(x.shape)] + [bf(x, bf_arg) for bf_arg in bf_args], axis=1)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Straight line fitting\n", "\n", "For straight line fitting, we use a linear regression model of the form $y(x, \\mathbf{w}) = w_0 + w_1 x$ and do Bayesian inference for model parameters $\\mathbf{w}$. Predictions are made with the posterior predictive distribution. Since this model has only two parameters, $w_0$ and $w_1$, we can visualize the posterior density in 2D which is done in the first column of the following output. Rows use an increasing number of training data from a training dataset." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "from bayesian_linear_regression_util import *\n", "\n", "import matplotlib.pyplot as plt\n", "%matplotlib inline\n", "\n", "# Training dataset sizes\n", "N_list = [1, 3, 20]\n", "\n", "beta = 25.0\n", "alpha = 2.0\n", "\n", "# Training observations in [-1, 1)\n", "X = np.random.rand(N_list[-1], 1) * 2 - 1\n", "\n", "# Training target values\n", "t = f(X, noise_variance=1/beta)\n", "\n", "# Test observations\n", "X_test = np.linspace(-1, 1, 100).reshape(-1, 1)\n", "\n", "# Function values without noise \n", "y_true = f(X_test, noise_variance=0)\n", " \n", "# Design matrix of test observations\n", "Phi_test = expand(X_test, identity_basis_function)\n", "\n", "plt.figure(figsize=(15, 10))\n", "plt.subplots_adjust(hspace=0.4)\n", "\n", "for i, N in enumerate(N_list):\n", " X_N = X[:N]\n", " t_N = t[:N]\n", "\n", " # Design matrix of training observations\n", " Phi_N = expand(X_N, identity_basis_function)\n", " \n", " # Mean and covariance matrix of posterior\n", " m_N, S_N = posterior(Phi_N, t_N, alpha, beta)\n", " \n", " # Mean and variances of posterior predictive \n", " y, y_var = posterior_predictive(Phi_test, m_N, S_N, beta)\n", " \n", " # Draw 5 random weight samples from posterior and compute y values\n", " w_samples = np.random.multivariate_normal(m_N.ravel(), S_N, 5).T\n", " y_samples = Phi_test.dot(w_samples)\n", " \n", " plt.subplot(len(N_list), 3, i * 3 + 1)\n", " plot_posterior(m_N, S_N, f_w0, f_w1)\n", " plt.title(f'Posterior density (N = {N})')\n", " plt.legend()\n", "\n", " plt.subplot(len(N_list), 3, i * 3 + 2)\n", " plot_data(X_N, t_N)\n", " plot_truth(X_test, y_true)\n", " plot_posterior_samples(X_test, y_samples)\n", " plt.ylim(-1.5, 1.0)\n", " plt.legend()\n", "\n", " plt.subplot(len(N_list), 3, i * 3 + 3)\n", " plot_data(X_N, t_N)\n", " plot_truth(X_test, y_true, label=None)\n", " plot_predictive(X_test, y, np.sqrt(y_var))\n", " plt.ylim(-1.5, 1.0)\n", " plt.legend()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In the second column, 5 random weight samples are drawn from the posterior and the corresponding regression lines are plotted in red color. The line resulting from the true parameters, `f_w0` and `f_w1` is plotted as dashed black line and the noisy training data as black dots. The third column shows the mean and the standard deviation of the posterior predictive distribution along with the true model and the training data.\n", "\n", "It can be clearly seen how the posterior density in the first column gets more sharply peaked as the size of the dataset increases which corresponds to a decrease in the sample variance in the second column and to a decrease in prediction uncertainty as shown in the third column. Also note how prediction uncertainty is higher in regions of less observations. \n", "\n", "### Gaussian basis functions\n", "\n", "The following example demonstrates how to fit a Gaussian basis function model to a noisy sinusoidal dataset. It uses 9 Gaussian basis functions with mean values equally distributed over $[0, 1]$ each having a standard deviation of $0.1$. Inference for parameters $\\mathbf{w}$ is done in the same way as in the previous example except that we now infer values for 10 parameters (bias term $w_0$ and $w_1,\\ldots,w_9$ for the 9 basis functions) instead of 2. We therefore cannot display the posterior density unless we selected 2 parameters at random." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "N_list = [3, 8, 20]\n", "\n", "beta = 25.0\n", "alpha = 2.0\n", "\n", "# Training observations in [-1, 1)\n", "X = np.random.rand(N_list[-1], 1)\n", "\n", "# Training target values\n", "t = g(X, noise_variance=1/beta)\n", "\n", "# Test observations\n", "X_test = np.linspace(0, 1, 100).reshape(-1, 1)\n", "\n", "# Function values without noise \n", "y_true = g(X_test, noise_variance=0)\n", " \n", "# Design matrix of test observations\n", "Phi_test = expand(X_test, bf=gaussian_basis_function, bf_args=np.linspace(0, 1, 9))\n", "\n", "plt.figure(figsize=(10, 10))\n", "plt.subplots_adjust(hspace=0.4)\n", "\n", "for i, N in enumerate(N_list):\n", " X_N = X[:N]\n", " t_N = t[:N]\n", "\n", " # Design matrix of training observations\n", " Phi_N = expand(X_N, bf=gaussian_basis_function, bf_args=np.linspace(0, 1, 9))\n", "\n", " # Mean and covariance matrix of posterior\n", " m_N, S_N = posterior(Phi_N, t_N, alpha, beta)\n", " \n", " # Mean and variances of posterior predictive \n", " y, y_var = posterior_predictive(Phi_test, m_N, S_N, beta)\n", " \n", " # Draw 5 random weight samples from posterior and compute y values\n", " w_samples = np.random.multivariate_normal(m_N.ravel(), S_N, 5).T\n", " y_samples = Phi_test.dot(w_samples)\n", " \n", " plt.subplot(len(N_list), 2, i * 2 + 1)\n", " plot_data(X_N, t_N)\n", " plot_truth(X_test, y_true)\n", " plot_posterior_samples(X_test, y_samples)\n", " plt.ylim(-1.0, 2.0)\n", " plt.legend()\n", " \n", " plt.subplot(len(N_list), 2, i * 2 + 2)\n", " plot_data(X_N, t_N)\n", " plot_truth(X_test, y_true, label=None)\n", " plot_predictive(X_test, y, np.sqrt(y_var))\n", " plt.ylim(-1.0, 2.0)\n", " plt.legend()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Again, as the size of the dataset increases the posterior sample variance and the prediction uncertainty decreases. Also, regions with less observations have higher prediction uncertainty.\n", "\n", "### Evidence evaluation\n", "\n", "As already mentioned, the evidence function or marginal likelihood can be used to compare models of different complexity using training data alone. This is shown here for 10 polynomial basis function models of different degree using a sinusoidal dataset generated with `g`. For evaluating the log marginal likelihood we implement $(18)$ as `log_marginal_likelihood` function. " ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "def log_marginal_likelihood(Phi, t, alpha, beta):\n", " \"\"\"Computes the log of the marginal likelihood.\"\"\"\n", " N, M = Phi.shape\n", "\n", " m_N, _, S_N_inv = posterior(Phi, t, alpha, beta, return_inverse=True)\n", " \n", " E_D = beta * np.sum((t - Phi.dot(m_N)) ** 2)\n", " E_W = alpha * np.sum(m_N ** 2)\n", " \n", " score = M * np.log(alpha) + \\\n", " N * np.log(beta) - \\\n", " E_D - \\\n", " E_W - \\\n", " np.log(np.linalg.det(S_N_inv)) - \\\n", " N * np.log(2 * np.pi)\n", "\n", " return 0.5 * score" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The 10 polynomial basis function models of degrees 0-9 are compared based on the log marginal likelihood computed with a dataset of 10 observations. We still assume that the values of $\\alpha$ and $\\beta$ are known and will see in the next section how they can be inferred by maximizing the log marginal likelihood. When plotting the posterior predictive distribution of the polynomial models we can see that a model of degree 3 has already sufficient complexity to explain the data reasonably well." ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "N = 10\n", "\n", "beta = 1 / (0.3 ** 2)\n", "alpha = 0.005\n", "\n", "degree = 9\n", "degrees = range(degree + 1)\n", "\n", "X = np.linspace(0, 1, N).reshape(-1, 1)\n", "t = g(X, noise_variance=1/beta)\n", "\n", "Phi = expand(X, bf=polynomial_basis_function, bf_args=degrees[1:])\n", "Phi_test = expand(X_test, bf=polynomial_basis_function, bf_args=degrees[1:])\n", "\n", "plt.figure(figsize=(18, 8))\n", "plt.subplots_adjust(hspace=0.4)\n", "\n", "for d in degrees:\n", " up = d + 1\n", " m_N, S_N = posterior(Phi[:,:up], t, alpha, beta)\n", " y, y_var = posterior_predictive(Phi_test[:,:up], m_N, S_N, beta)\n", "\n", " plt.subplot(2, 5, up)\n", " plot_data(X, t)\n", " plot_truth(X_test, y_true, label=None)\n", " plot_predictive(X_test, y, np.sqrt(y_var), y_label=None, std_label=None, plot_xy_labels=False)\n", " plt.title(f'Degree = {d}')\n", " plt.ylim(-1.0, 2.0)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We also see how polynomial models of higher degree do not overfit to the dataset which is a consequence of using a prior over model parameters that favors small(er) parameter values. This is equivalent to minimizing a sum-of-squares error function plus a quadratic regularization term whose strength is given by ratio $\\alpha / \\beta$ as can be seen from equation $(12)$. \n", "\n", "When evaluating the log marginal likelihood for all 10 polynomial models we usually obtain the highest value for models of degree 3 or 4 (depending on the non-deterministic part i.e. noise of the generated dataset results may vary slightly). This is consistent with the observation that a polynomial model of degree 3 already explains the data sufficiently well and confirms that marginal likelihood evaluation favors models of intermediate complexity." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "mlls = []\n", "\n", "for d in degrees:\n", " mll = log_marginal_likelihood(Phi[:,:d+1], t, alpha=alpha, beta=beta)\n", " mlls.append(mll)\n", "\n", "degree_max = np.argmax(mlls)\n", " \n", "plt.plot(degrees, mlls)\n", "plt.axvline(x=degree_max, ls='--', c='k', lw=1)\n", "plt.xticks(range(0, 10))\n", "plt.xlabel('Polynomial degree')\n", "plt.ylabel('Log marginal likelihood');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "It is also interesting to see that a polynomial model of degree 1 (straight line) seems to explain the data better than a model of degree 2. This is because the data-generating sinusoidal function has no even terms in a polynomial expansion. A model of degree 2 therefore only adds complexity without being able to explain the data better. This higher complexity is penalized by the evidence function (see also section 3.4. in [PRML](https://www.microsoft.com/en-us/research/uploads/prod/2006/01/Bishop-Pattern-Recognition-and-Machine-Learning-2006.pdf))." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evidence maximization" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "So far we have assumed that values of $\\alpha$ and $\\beta$ are known. In most situations however, they are unknown and must be inferred. Iterating over equations $(21)$ and $(22)$ until convergence jointly infers the posterior distribution over parameters $\\mathbf{w}$ and optimal values for parameters $\\alpha$ and $\\beta$. This is implemented in the following `fit` function. We start with small values for $\\alpha$ and $\\beta$ corresponding to a low precision (= high variance) of prior $(8)$ and conditional density $(3)$." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "def fit(Phi, t, alpha_0=1e-5, beta_0=1e-5, max_iter=200, rtol=1e-5, verbose=False):\n", " \"\"\"\n", " Jointly infers the posterior sufficient statistics and optimal values \n", " for alpha and beta by maximizing the log marginal likelihood.\n", " \n", " Args:\n", " Phi: Design matrix (N x M).\n", " t: Target value array (N x 1).\n", " alpha_0: Initial value for alpha.\n", " beta_0: Initial value for beta.\n", " max_iter: Maximum number of iterations.\n", " rtol: Convergence criterion.\n", " \n", " Returns:\n", " alpha, beta, posterior mean, posterior covariance.\n", " \"\"\"\n", " \n", " N, M = Phi.shape\n", "\n", " eigenvalues_0 = np.linalg.eigvalsh(Phi.T.dot(Phi))\n", "\n", " beta = beta_0\n", " alpha = alpha_0\n", "\n", " for i in range(max_iter):\n", " beta_prev = beta\n", " alpha_prev = alpha\n", "\n", " eigenvalues = eigenvalues_0 * beta\n", "\n", " m_N, S_N, S_N_inv = posterior(Phi, t, alpha, beta, return_inverse=True)\n", "\n", " gamma = np.sum(eigenvalues / (eigenvalues + alpha))\n", " alpha = gamma / np.sum(m_N ** 2)\n", "\n", " beta_inv = 1 / (N - gamma) * np.sum((t - Phi.dot(m_N)) ** 2)\n", " beta = 1 / beta_inv\n", "\n", " if np.isclose(alpha_prev, alpha, rtol=rtol) and np.isclose(beta_prev, beta, rtol=rtol):\n", " if verbose:\n", " print(f'Convergence after {i + 1} iterations.')\n", " return alpha, beta, m_N, S_N\n", "\n", " if verbose:\n", " print(f'Stopped after {max_iter} iterations.')\n", " return alpha, beta, m_N, S_N" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We now generate a sinusoidal training dataset of size 30 with variance $\\beta^{-1} = 0.3^2$ and then use `fit` to obtain the posterior over parameters $\\mathbf{w}$ and optimal values for $\\alpha$ and $\\beta$. The used regression model is a polynomial model of degree 4." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Convergence after 18 iterations.\n" ] } ], "source": [ "N = 30\n", "\n", "degree = 4\n", "\n", "X = np.linspace(0, 1, N).reshape(-1, 1)\n", "t = g(X, noise_variance=0.3 ** 2)\n", "\n", "Phi = expand(X, bf=polynomial_basis_function, bf_args=range(1, degree + 1))\n", "\n", "alpha, beta, m_N, S_N = fit(Phi, t, rtol=1e-5, verbose=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Alternatively, we can also use [`BayesianRidge`](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.BayesianRidge.html#sklearn.linear_model.BayesianRidge) from scikit-learn for Bayesian regression. The `fit` and `predict` methods of this estimator are on the same abstraction level as our `fit` and `posterior_predictive` functions. The implementation of `BayesianRidge` is very similar to our implementation except that it uses [Gamma](https://en.wikipedia.org/wiki/Gamma_distribution) priors over parameters $\\alpha$ and $\\beta$. The default hyper-parameter values of the Gamma priors assign high probability density to low values for $\\alpha$ and $\\beta$. In our implementation, we simply start optimization from low $\\alpha$ and $\\beta$ values. Another difference is that `BayesianRidge` uses different parameter names (`lambda` instead of `alpha` and `alpha` instead of `beta`, see also section [Bayesian Regression](https://scikit-learn.org/stable/modules/linear_model.html#bayesian-regression) in the scikit-learn user guide)." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Convergence after 18 iterations\n" ] } ], "source": [ "from sklearn.linear_model import BayesianRidge\n", "\n", "br = BayesianRidge(fit_intercept=False, tol=1e-5, verbose=True)\n", "br.fit(Phi, t.ravel());" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "When comparing the results from our implementation with those from `BayesianRidge` we see that they are almost identical. In the following, inferred values for $\\alpha$, $\\beta$ and $\\mathbf{m}_N$ are compared as well as predictions and their uncertainties. Results prefixed with `np` are those from our implementation, results prefixed with `br` are those obtained with `BayesianRidge`." ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Alpha\n", "-----\n", "np: 0.00743863611678815\n", "br: 0.007438637616154343\n", "\n" ] } ], "source": [ "print_comparison('Alpha', alpha, br.lambda_)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Beta\n", "----\n", "np: 10.326183677121568\n", "br: 10.326176707792914\n", "\n" ] } ], "source": [ "print_comparison('Beta', beta, br.alpha_)" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Weights\n", "-------\n", "np: [ 0.44679985 8.20305327 -17.8307425 -2.50715196 12.54716275]\n", "br: [ 0.44679957 8.20305668 -17.83075117 -2.50714534 12.54716184]\n", "\n" ] } ], "source": [ "print_comparison('Weights', m_N.ravel(), br.coef_)" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "# Test values at x = 0.3 and x = 0.7 \n", "X_test = np.array([[0.3], [0.7]])\n", "\n", "# Design matrix of test values\n", "Phi_test = expand(X_test, bf=polynomial_basis_function, bf_args=range(1, degree + 1))" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Prediction mean\n", "---------------\n", "np: [ 1.33688792 -0.39550603]\n", "br: [ 1.33688805 -0.39550613]\n", "\n", "Prediction std\n", "--------------\n", "np: [0.3262517 0.32725117]\n", "br: [0.3262518 0.32725127]\n", "\n" ] } ], "source": [ "y_np_mean, y_np_var = posterior_predictive(Phi_test, m_N, S_N, beta)\n", "y_br_mean, y_br_std = br.predict(Phi_test, return_std=True)\n", "\n", "print_comparison('Prediction mean', y_np_mean.ravel(), y_br_mean)\n", "print_comparison('Prediction std', np.sqrt(y_np_var), y_br_std)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "An alternative, non-parametric approach to Bayesian regression are [Gaussian processes](http://krasserm.github.io/2018/03/19/gaussian-processes/) which infer distributions over functions directly instead of distributions over parameters of parametric models." ] } ], "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": "3.6.8" } }, "nbformat": 4, "nbformat_minor": 2 }