{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# MGC-X and DCorr-X: Independence Testing for Time Series" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this tutorial, we explore\n", "\n", "- The theory behind the Cross Distance Correlation (DCorr-X) and Cross Multiscale Graph Correlation (MGC-X) tests\n", "- The unique methodological features such as optimal scale and optimal lag\n", "- The features of the implementation" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Theory" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Notation\n", "Let $\\mathbb{N}$ be the non-negative integers $\\{0, 1, 2, ...\\}$, and $\\mathbb{R}$ be the real line $(-\\infty, \\infty)$. Let $F_X$, $F_Y$, and $F_{X,Y}$ represent the marginal and joint distributions of random variables $X$ and $Y$, whose realizations exist in $\\mathcal{X}$ and $\\mathcal{Y}$, respectively. Similarly, Let $F_{X_t}$, $F_{Y_s}$, and $F_{(X_t,Y_s)}$ represent the marginal and joint distributions of the time-indexed random variables $X_t$ and $Y_s$ at timesteps $t$ and $s$. For this work, assume $\\mathcal{X} = \\mathbb{R}^p$ and $\\mathcal{Y} = \\mathbb{R}^q$ for $p, q > 0$. Finally, let $\\{(X_t,Y_t)\\}_{t=-\\infty}^{\\infty}$ represent the full, jointly-sampled time series, structured as a countably long list of observations $(X_t, Y_t)$. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Problem Statement\n", "The test addresses the problem of independence testing for time series. To formalize the problem, consider a strictly stationary time series $\\{(X_t,Y_t)\\}_{t=-\\infty}^{\\infty}$, with the observed sample $\\{(X_1,Y_1),...,(X_n, Y_n)\\}$. Choose some $M \\in \\mathbb{N}$, the maximum_lag hyperparameter. We test the independence of two series via the following hypothesis.\n", "\n", "\\begin{align*}\n", " H_0: F_{(X_t,Y_{t-j})} &= F_{X_t} F_{Y_{t-j}} \\text{ for each } j \\in \\{0, 1, ..., M\\}\\\\\n", " H_A: F_{(X_t,Y_{t-j})} &\\neq F_{X_t} F_{Y_{t-j}} \\text{ for some } j \\in \\{0, 1, ..., M\\}\n", "\\end{align*}\n", "\n", "The null hypothesis implies that for any $(M+1)$-length stretch in the time series, $X_t$ is pairwise independent of present and past values $Y_{t-j}$ spaced $j$ timesteps away (including $j=0$). A corresponding test for whether $Y_t$ is dependent on past values of $X_t$ is available by swapping the labels of each time series. Finally, the hyperparameter $M$ governs the maximum number of timesteps in the past for which we check the influence of $Y_{t-j}$ on $X_t$. This $M$ can be chosen for computation considerations, as well as for specific subject matter purposes, e.g. a signal from one region of the brain might only influence be able to influence another within 20 time steps implies $M = 20$. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### The Test Statistic\n", "Define the **cross-distance correlation** at lag $j$ as\n", "\n", "\\begin{align*}\n", " \\text{DCorr}(j) := \\text{DCorr}(X_t, Y_{t-j}).\n", "\\end{align*}\n", "\n", "Where $\\text{DCorr}(\\cdot, \\cdot)$ is the distance correlation function. Assuming strict stationarity of $\\{(X_t,Y_t)\\}$ is important in even defining $\\text{DCorr}(j)$, as the parameter depends only on the spacing $j$, and not the timestep $t$ of $X_t$ and $Y_{t-j}$. Similarly, let $\\text{DCorr}n(j)$ be its estimator, with $\\text{MGC}_n(j)$ being the $\\text{MGC}$ test statistic evaluated for $\\{X_t\\}$ and $\\{Y_{t-j}\\}$. The $\\text{DCorr-X}^M$ test statistic is \n", "\n", "\\begin{align*}\n", " \\text{DCorr-X}_n^M &= \\sum_{j=0}^{M} \\left(\\frac{n-j}{n}\\right) \\cdot \\text{DCorr}n(j).\n", "\\end{align*}\n", "\n", "Similarly, the $\\text{MGC-X}$ test statistic is \n", "\n", "\\begin{align*}\n", " \\text{MGC-X}_n^M &= \\sum_{j=0}^{M} \\left(\\frac{n-j}{n}\\right) \\cdot \\text{MGC}_n(j).\n", "\\end{align*}\n", "\n", "While $\\text{MGC-X}$ is more computationally intensive than $\\text{DCorr-X}$, $\\text{MGC-X}$ employs multiscale analysis to achieve better finite-sample power in high-dimensional, nonlinear, and structured data settings [[1]](https://elifesciences.org/articles/41690)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### The P-Value\n", "Let $T_n$ represent either of the test statistics above. To compute the p-value, one need to estimate the null distribution of $T_n$, namely its distribution under indepdendence pair of data. A typical permutation test would permute the indices $\\{1,2,3,...,n\\}$, reorder the series $\\{Y_t\\}$ according to this permutation, and $T_n$ would be computed on $\\{X_t\\}$ and the reordered $\\{Y_t\\}$. This procedure would be repeated $K$ times, generating $K$ samples of the test statistic under the null. This permutation test requires exchangeability of the sequence $\\{Y_t\\}$, which would be true in the i.i.d. case, but is generally violated in the time series case. Instead, a block permutation captures the dependence between elements of the series, as described in \\cite{politis2003}. Letting $\\lceil \\cdot \\rceil$ be the ceiling function, this procedure partitions the list of indices into size $b$ \"blocks\", and permutes the $\\lceil \\frac{n}{b} \\rceil$ blocks in order to generate samples of the test statistic under the null.\n", "Specifically,\n", "\n", "1. Choose a random permutation of the indices $\\{0, 1, 2, ..., \\lceil \\frac{n}{b} \\rceil\\}$. \n", "\n", "2. From index $i$ in the permutation, produce block $B_{i} = (Y_{bi+1},Y_{bi+2},...,Y_{bi + b})$, which is a section of the series $\\{Y_t\\}$.\n", "\n", "3. Let the series $\\{Y_{\\pi(1)}, ..., Y_{\\pi(n)}\\} = (B_1, B_2, ..., B_{\\frac{n}{b}})$, where $\\pi$ maps indices $\\{1,2,...,n\\}$ to the new, block permuted indices.\n", "\n", "4. Compute $T^{(r)}_n$ on the series $\\{(X_t, Y_{\\pi(t)})\\}_{t=1}^n$ for replicate $r$.\n", "\n", "Repeat this procedure $K$ times (typically $K = 100$ or $1000$), and let $T^{(0)}_n = T_n$, with:\n", "\n", "\\begin{align*}\n", " p\\text{-value}(T_n) &= \\frac{1}{K+1} \\sum_{r=0}^K \\mathbb{I}\\{T^{(r)}_n \\geq T_n\\}\n", "\\end{align*}\n", "\n", "where $\\mathbb{I}\\{\\cdot\\}$ is the indicator function." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Using DCorr-X and MGC-X" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "%matplotlib inline\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import random\n", "from scipy.stats import pearsonr\n", "\n", "from mgcpy.independence_tests.dcorrx import DCorrX\n", "from mgcpy.independence_tests.mgcx import MGCX" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Simulate time series\n", "Let $\\epsilon_t$ and $\\eta_t$ be i.i.d. standard normally distributed." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Independent AR(1):\n", "$$\\begin{bmatrix}\n", " X_t\\\\\n", " Y_t\n", " \\end{bmatrix}\n", " =\n", " \\begin{bmatrix}\n", " 0.5 & 0\\\\\n", " 0 & 0.5\n", " \\end{bmatrix}\n", " \\begin{bmatrix}\n", " X_{t-1}\\\\\n", " Y_{t-1}\n", " \\end{bmatrix}\n", " +\n", " \\begin{bmatrix}\n", " \\epsilon_t\\\\\n", " \\eta_t\n", " \\end{bmatrix}$$" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "def indep_ar1(n, phi = 0.5, sigma2 = 1.0):\n", " # X_t and Y_t are univarite AR(1) with phi = 0.5 for both series.\n", " # Noise follows N(0, sigma2).\n", " \n", " # Innovations.\n", " epsilons = np.random.normal(0.0, sigma2, n)\n", " etas = np.random.normal(0.0, sigma2, n)\n", " \n", " X = np.zeros(n)\n", " Y = np.zeros(n)\n", " X[0] = epsilons[0]\n", " Y[0] = etas[0]\n", " \n", " # AR(1) process.\n", " for t in range(1,n):\n", " X[t] = phi*X[t-1] + epsilons[t]\n", " Y[t] = phi*Y[t-1] + etas[t]\n", " \n", " return X, Y" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Crosscorrelated AR(1):\n", "$$\\begin{bmatrix}\n", " X_t\\\\\n", " Y_t\n", " \\end{bmatrix}\n", " =\n", " \\begin{bmatrix}\n", " 0 & 0.5\\\\\n", " 0.5 & 0\n", " \\end{bmatrix}\n", " \\begin{bmatrix}\n", " X_{t-1}\\\\\n", " Y_{t-1}\n", " \\end{bmatrix}\n", " +\n", " \\begin{bmatrix}\n", " \\epsilon_t\\\\\n", " \\eta_t\n", " \\end{bmatrix}$$" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "def cross_corr_ar1(n, phi = 0.5, sigma2 = 1.0):\n", " # X_t and Y_t are together a bivarite AR(1) with Phi = [0 0.5; 0.5 0].\n", " # Innovations follow N(0, sigma2).\n", " \n", " # Innovations.\n", " epsilons = np.random.normal(0.0, sigma2, n)\n", " etas = np.random.normal(0.0, sigma2, n)\n", " \n", " X = np.zeros(n)\n", " Y = np.zeros(n)\n", " X[0] = epsilons[0]\n", " Y[0] = etas[0]\n", "\n", " for t in range(1,n):\n", " X[t] = phi*Y[t-1] + epsilons[t]\n", " Y[t] = phi*X[t-1] + etas[t]\n", " \n", " return X, Y" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Nonlinearly related at lag 1:\n", "$$\\begin{bmatrix}\n", " X_t\\\\\n", " Y_t\n", " \\end{bmatrix}\n", " =\n", " \\begin{bmatrix}\n", " \\epsilon_t Y_{t-1}\\\\\n", " \\eta_t\n", " \\end{bmatrix}$$" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "def nonlinear_lag1(n, phi = 1, sigma2 = 1):\n", " # X_t and Y_t are together a bivarite nonlinear process.\n", " # Innovations follow N(0, sigma2).\n", " \n", " # Innovations.\n", " epsilons = np.random.normal(0.0, sigma2, n)\n", " etas = np.random.normal(0.0, sigma2, n)\n", " \n", " X = np.zeros(n)\n", " Y = np.zeros(n)\n", " Y[0] = etas[0]\n", " \n", " for t in range(1,n):\n", " X[t] = phi*epsilons[t]*Y[t-1]\n", " Y[t] = etas[t]\n", " \n", " return X, Y" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Plot time series" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "def plot_ts(X, Y, title, xlab = \"X_t\", ylab = \"Y_t\"):\n", " n = X.shape[0]\n", " t = range(1, n + 1)\n", " \n", " fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(15,7.5))\n", " fig.suptitle(title)\n", " plt.rcParams.update({'font.size': 15})\n", " \n", " ax1.plot(t, X)\n", " ax1.plot(t, Y)\n", " ax1.legend(['X_t', 'Y_t'], loc = 'upper left', prop={'size': 12})\n", " ax1.set_xlabel(\"t\")\n", " \n", " ax2.scatter(X,Y, color=\"black\") \n", " ax2.set_ylabel(ylab)\n", " ax2.set_xlabel(xlab)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Explore with DCorr-X and MGC-X." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "def compute_dcorrx(X, Y, max_lag):\n", " dcorrx = DCorrX(max_lag = max_lag, which_test = 'unbiased')\n", " dcorrx_statistic, metadata = dcorrx.test_statistic(X, Y)\n", " p_value, _ = dcorrx.p_value(X, Y)\n", " optimal_lag = metadata['optimal_lag']\n", "\n", " print(\"DCorrX test statistic:\", dcorrx_statistic)\n", " print(\"P Value:\", p_value)\n", " print(\"Optimal Lag:\", optimal_lag)\n", "\n", "def compute_mgcx(X, Y, max_lag):\n", " mgcx = MGCX(max_lag = max_lag)\n", " mgcx_statistic, metadata = mgcx.test_statistic(X, Y)\n", " p_value, _ = mgcx.p_value(X, Y)\n", " optimal_lag = metadata['optimal_lag']\n", " optimal_scale = metadata['optimal_scale']\n", " \n", " print(\"MGCX test statistic:\", mgcx_statistic)\n", " print(\"P Value:\", p_value)\n", " print(\"Optimal Lag:\", optimal_lag)\n", " print(\"Optimal Scale:\", optimal_scale)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "DCorrX test statistic: 0.0\n", "P Value: 0.418\n", "Optimal Lag: 0\n", "MGCX test statistic: 0.0\n", "P Value: 0.424\n", "Optimal Lag: 0\n", "Optimal Scale: [40, 40]\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "