{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## State space models - Chandrasekhar recursions" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "execution": { "iopub.execute_input": "2021-02-02T06:57:43.332939Z", "iopub.status.busy": "2021-02-02T06:57:43.332170Z", "iopub.status.idle": "2021-02-02T06:57:47.535510Z", "shell.execute_reply": "2021-02-02T06:57:47.536246Z" } }, "outputs": [], "source": [ "%matplotlib inline\n", "\n", "import numpy as np\n", "import pandas as pd\n", "import statsmodels.api as sm\n", "import matplotlib.pyplot as plt\n", "\n", "from pandas_datareader.data import DataReader" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Although most operations related to state space models rely on the Kalman filtering recursions, in some special cases one can use a separate method often called \"Chandrasekhar recursions\". These provide an alternative way to iteratively compute the conditional moments of the state vector, and in some cases they can be substantially less computationally intensive than the Kalman filter recursions. For complete details, see the paper \"Using the 'Chandrasekhar Recursions' for Likelihood Evaluation of DSGE Models\" (Herbst, 2015). Here we just sketch the basic idea." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### State space models and the Kalman filter\n", "\n", "Recall that a time-invariant state space model can be written:\n", "\n", "$$\n", "\\begin{aligned}\n", "y_t &= Z \\alpha_t + \\varepsilon_t, \\qquad \\varepsilon_t \\sim N(0, H) \\\\\n", "\\alpha_{t+1} & = T \\alpha_t + R \\eta_t, \\qquad \\eta_t \\sim N(0, Q) \\\\\n", "\\alpha_1 & \\sim N(a_1, P_1)\n", "\\end{aligned}\n", "$$\n", "\n", "where $y_t$ is a $p \\times 1$ vector and $\\alpha_t$ is an $m \\times 1$ vector.\n", "\n", "Each iteration of the Kalman filter, say at time $t$, can be split into three parts:\n", "\n", "1. **Initialization**: specification of $a_t$ and $P_t$ that define the conditional state distribution, $\\alpha_t \\mid y^{t-1} \\sim N(a_t, P_t)$.\n", "2. **Updating**: computation of $a_{t|t}$ and $P_{t|t}$ that define the conditional state distribution, $\\alpha_t \\mid y^{t} \\sim N(a_{t|t}, P_{t|t})$.\n", "3. **Prediction**: computation of $a_{t+1}$ and $P_{t+1}$ that define the conditional state distribution, $\\alpha_{t+1} \\mid y^{t} \\sim N(a_{t+1}, P_{t+1})$.\n", "\n", "Of course after the first iteration, the prediction part supplies the values required for initialization of the next step." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Focusing on the prediction step, the Kalman filter recursions yield:\n", "\n", "$$\n", "\\begin{aligned}\n", "a_{t+1} & = T a_{t|t} \\\\\n", "P_{t+1} & = T P_{t|t} T' + R Q R' \\\\\n", "\\end{aligned}\n", "$$\n", "\n", "where the matrices $T$ and $P_{t|t}$ are each $m \\times m$, where $m$ is the size of the state vector $\\alpha$. In some cases, the state vector can become extremely large, which can imply that the matrix multiplications required to produce $P_{t+1}$ can be become computationally intensive." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Example: seasonal autoregression\n", "\n", "As an example, notice that an AR(r) model (we use $r$ here since we already used $p$ as the dimension of the observation vector) can be put into state space form as:\n", "\n", "$$\n", "\\begin{aligned}\n", "y_t &= \\alpha_t \\\\\n", "\\alpha_{t+1} & = T \\alpha_t + R \\eta_t, \\qquad \\eta_t \\sim N(0, Q)\n", "\\end{aligned}\n", "$$\n", "\n", "where:\n", "\n", "\n", "$$\n", "\\begin{aligned}\n", "T = \\begin{bmatrix}\n", "\\phi_1 & \\phi_2 & \\dots & \\phi_r \\\\\n", "1 & 0 & & 0 \\\\\n", "\\vdots & \\ddots & & \\vdots \\\\\n", "0 & & 1 & 0 \\\\\n", "\\end{bmatrix} \\qquad\n", "R = \\begin{bmatrix}\n", "1 \\\\\n", "0 \\\\\n", "\\vdots \\\\\n", "0\n", "\\end{bmatrix} \\qquad\n", "Q = \\begin{bmatrix}\n", "\\sigma^2\n", "\\end{bmatrix}\n", "\\end{aligned}\n", "$$\n", "\n", "In an AR model with daily data that exhibits annual seasonality, we might want to fit a model that incorporates lags up to $r=365$, in which case the state vector would be at least $m = 365$. The matrices $T$ and $P_{t|t}$ then each have $365^2 = 133225$ elements, and so most of the time spent computing the likelihood function (via the Kalman filter) can become dominated by the matrix multiplications in the prediction step." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### State space models and the Chandrasekhar recursions\n", "\n", "The Chandrasekhar recursions replace equation $P_{t+1} = T P_{t|t} T' + R Q R'$ with a different recursion:\n", "\n", "$$\n", "P_{t+1} = P_t + W_t M_t W_t'\n", "$$\n", "\n", "but where $W_t$ is a matrix with dimension $m \\times p$ and $M_t$ is a matrix with dimension $p \\times p$, where $p$ is the dimension of the observed vector $y_t$. These matrices themselves have recursive formulations. For more general details and for the formulas for computing $W_t$ and $M_t$, see Herbst (2015).\n", "\n", "**Important note**: unlike the Kalman filter, the Chandrasekhar recursions can not be used for every state space model. In particular, the latter has the following restrictions (that are not required for the use of the former):\n", "\n", "- The model must be time-invariant, except that time-varying intercepts are permitted.\n", "- Stationary initialization of the state vector must be used (this rules out all models in non-stationary components)\n", "- Missing data is not permitted" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To understand why this formula can imply more efficient computations, consider again the SARIMAX case, above. In this case, $p = 1$, so that $M_t$ is a scalar and we can rewrite the Chandrasekhar recursion as:\n", "\n", "$$\n", "P_{t+1} = P_t + M_t \\times W_t W_t'\n", "$$\n", "\n", "The matrices being multiplied, $W_t$, are then of dimension $m \\times 1$, and in the case $r=365$, they each only have $365$ elements, rather than $365^2$ elements. This implies substantially fewer computations are required to complete the prediction step." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Convergence\n", "\n", "A factor that complicates a straightforward discussion of performance implications is the well-known fact that in time-invariant models, the predicted state covariance matrix will converge to a constant matrix. This implies that there exists an $S$ such that, for every $t > S$, $P_t = P_{t+1}$. Once convergence has been achieved, we can eliminate the equation for $P_{t+1}$ from the prediction step altogether.\n", "\n", "In simple time series models, like AR(r) models, convergence is achieved fairly quickly, and this can limit the performance benefit to using the Chandrasekhar recursions. Herbst (2015) focuses instead on DSGE (Dynamic Stochastic General Equilibrium) models instead, which often have a large state vector and often a large number of periods to achieve convergence. In these cases, the performance gains can be quite substantial." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Practical example\n", "\n", "As a practical example, we will consider monthly data that has a clear seasonal component. In this case, we look at the inflation rate of apparel, as measured by the consumer price index. A graph of the data indicates strong seasonality." ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "execution": { "iopub.execute_input": "2021-02-02T06:57:47.545102Z", "iopub.status.busy": "2021-02-02T06:57:47.544272Z", "iopub.status.idle": "2021-02-02T06:57:48.851432Z", "shell.execute_reply": "2021-02-02T06:57:48.852188Z" } }, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "cpi_apparel = DataReader('CPIAPPNS', 'fred', start='1986')\n", "cpi_apparel.index = pd.DatetimeIndex(cpi_apparel.index, freq='MS')\n", "inf_apparel = np.log(cpi_apparel).diff().iloc[1:] * 1200\n", "inf_apparel.plot(figsize=(15, 5));" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will construct two model instances. The first will be set to use the Kalman filter recursions, while the second will be set to use the Chandrasekhar recursions. This setting is controlled by the `ssm.filter_chandrasekhar` property, as shown below.\n", "\n", "The model we have in mind is a seasonal autoregression, where we include the first 6 months as lags as well as the given month in each of the previous 15 years as lags. This implies that the state vector has dimension $m = 186$, which is large enough that we might expect to see some substantial performance gains by using the Chandrasekhar recursions.\n", "\n", "**Remark**: We set `tolerance=0` in each model - this has the effect of preventing the filter from ever recognizing that the prediction covariance matrix has converged. *This is not recommended in practice*. We do this here to highlight the superior performance of the Chandrasekhar recursions when they are used in every period instead of the typical Kalman filter recursions. Later, we will show the performance in a more realistic setting that we do allow for convergence." ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "execution": { "iopub.execute_input": "2021-02-02T06:57:48.866271Z", "iopub.status.busy": "2021-02-02T06:57:48.865399Z", "iopub.status.idle": "2021-02-02T06:57:48.895699Z", "shell.execute_reply": "2021-02-02T06:57:48.896380Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "186\n" ] } ], "source": [ "# Model that will apply Kalman filter recursions\n", "mod_kf = sm.tsa.SARIMAX(inf_apparel, order=(6, 0, 0), seasonal_order=(15, 0, 0, 12), tolerance=0)\n", "print(mod_kf.k_states)\n", "\n", "# Model that will apply Chandrasekhar recursions\n", "mod_ch = sm.tsa.SARIMAX(inf_apparel, order=(6, 0, 0), seasonal_order=(15, 0, 0, 12), tolerance=0)\n", "mod_ch.ssm.filter_chandrasekhar = True" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We time computation of the log-likelihood function, using the following code:\n", "\n", "```python\n", "%timeit mod_kf.loglike(mod_kf.start_params)\n", "%timeit mod_ch.loglike(mod_ch.start_params)\n", "```\n", "\n", "This results in:\n", "\n", "```\n", "171 ms ± 19.7 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "85 ms ± 4.97 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "```\n", "\n", "The implication is that in this experiment, the Chandrasekhar recursions improved performance by about a factor of 2." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we mentioned above, in the previous experiment we disabled convergence of the predicted covariance matrices, so the results there are an upper bound. Now we allow for convergence, as usual, by removing the `tolerance=0` argument:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "execution": { "iopub.execute_input": "2021-02-02T06:57:48.905296Z", "iopub.status.busy": "2021-02-02T06:57:48.904489Z", "iopub.status.idle": "2021-02-02T06:57:48.943072Z", "shell.execute_reply": "2021-02-02T06:57:48.943721Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "186\n" ] } ], "source": [ "# Model that will apply Kalman filter recursions\n", "mod_kf = sm.tsa.SARIMAX(inf_apparel, order=(6, 0, 0), seasonal_order=(15, 0, 0, 12))\n", "print(mod_kf.k_states)\n", "\n", "# Model that will apply Chandrasekhar recursions\n", "mod_ch = sm.tsa.SARIMAX(inf_apparel, order=(6, 0, 0), seasonal_order=(15, 0, 0, 12))\n", "mod_ch.ssm.filter_chandrasekhar = True" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Again, we time computation of the log-likelihood function, using the following code:\n", "\n", "```python\n", "%timeit mod_kf.loglike(mod_kf.start_params)\n", "%timeit mod_ch.loglike(mod_ch.start_params)\n", "```\n", "\n", "This results in:\n", "\n", "```\n", "114 ms ± 7.64 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "70.5 ms ± 2.43 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "```\n", "\n", "The Chandrasekhar recursions still improve performance, but now only by about 33%. The reason for this is that after convergence, we no longer need to compute the predicted covariance matrices, so that for those post-convergence periods, there will be no difference in computation time between the two approaches. Below we check the period in which convergence was achieved:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "execution": { "iopub.execute_input": "2021-02-02T06:57:48.949175Z", "iopub.status.busy": "2021-02-02T06:57:48.948397Z", "iopub.status.idle": "2021-02-02T06:58:36.869961Z", "shell.execute_reply": "2021-02-02T06:58:36.871027Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Convergence at t=186, of T=419 total observations\n" ] } ], "source": [ "res_kf = mod_kf.filter(mod_kf.start_params)\n", "print('Convergence at t=%d, of T=%d total observations' %\n", " (res_kf.filter_results.period_converged, res_kf.nobs))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Since convergence happened relatively early, we are already avoiding the expensive matrix multiplications in more than half of the periods.\n", "\n", "However, as mentioned above, larger DSGE models may not achieve convergence for most or all of the periods in the sample, and so we could perhaps expect to achieve performance gains more similar to the first example. In their 2019 paper \"Euro area real-time density forecasting with financial or labor market frictions\", McAdam and Warne note that in their applications, \"Compared with the standard Kalman filter, it is our experience that these recursions speed up\n", "the calculation of the log-likelihood for the three models by roughly 50 percent\". This is about the same result as we found in our first example." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Aside on multithreaded matrix algebra routines\n", "\n", "The timings above are based on the Numpy installation installed via Anaconda, which uses Intel's MKL BLAS and LAPACK libraries. These implement multithreaded processing to speed up matrix algebra, which can be particularly helpful for operations on the larger matrices we're working with here. To get a sense of how this affects results, we could turn off multithreading by putting the following in the first cell of this notebook.\n", "\n", "```python\n", "import os\n", "os.environ[\"MKL_NUM_THREADS\"] = \"1\"\n", "```\n", "\n", "When we do this, the timings of the first example change to:\n", "\n", "```\n", "307 ms ± 3.08 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n", "97.5 ms ± 1.64 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "```\n", "\n", "and the timings of the second example change to:\n", "\n", "```\n", "178 ms ± 2.78 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n", "78.9 ms ± 950 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n", "```\n", "\n", "Both are slower, but the typical Kalman filter is affected much more.\n", "\n", "This is not unexpected; the performance differential between single and multithreaded linear algebra is much greater in the typical Kalman filter case, because the whole point of the Chandrasekhar recursions is to reduce the size of the matrix operations. It means that if multithreaded linear algebra is unavailable, the Chandrasekhar recursions offer even greater performance gains." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Chandrasekhar recursions and the univariate filtering approach\n", "\n", "It is also possible to combine the Chandrasekhar recursions with the univariate filtering approach of Koopman and Durbin (2000), by making use of the results of Aknouche and Hamdi in their 2007 paper \"Periodic Chandrasekhar recursions\". An initial implementation of this combination is included in Statsmodels. However, experiments suggest that this tends to degrade performance compared to even the usual Kalman filter. This accords with the computational savings reported for the univariate filtering method, which suggest that savings are highest when the state vector is small relative to the observation vector." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### Bibliography\n", "\n", "Aknouche, Abdelhakim, and Fayçal Hamdi. \"Periodic Chandrasekhar recursions.\" arXiv preprint arXiv:0711.3857 (2007).\n", "\n", "Herbst, Edward. \"Using the “Chandrasekhar Recursions” for likelihood evaluation of DSGE models.\" Computational Economics 45, no. 4 (2015): 693-705.\n", "\n", "Koopman, Siem J., and James Durbin. \"Fast filtering and smoothing for multivariate state space models.\" Journal of Time Series Analysis 21, no. 3 (2000): 281-296.\n", "\n", "McAdam, Peter, and Anders Warne. \"Euro area real-time density forecasting with financial or labor market frictions.\" International Journal of Forecasting 35, no. 2 (2019): 580-600." ] } ], "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": "3.7.9" } }, "nbformat": 4, "nbformat_minor": 2 }