diff --git a/examples/notebooks/wiggling.ipynb b/examples/notebooks/wiggling.ipynb index fa6b4aa..a0ff913 100644 --- a/examples/notebooks/wiggling.ipynb +++ b/examples/notebooks/wiggling.ipynb @@ -4,7 +4,7 @@ "metadata": { "colab": { "provenance": [], - "authorship_tag": "ABX9TyMi6wdHV1ohBr8hsQcGVJtj", + "authorship_tag": "ABX9TyMczfr4u3Vwpilu6STz6nX2", "include_colab_link": true }, "kernelspec": { @@ -53,8 +53,8 @@ { "cell_type": "markdown", "source": [ - "### Step 1: Choose a skater, and a wiggled version of the same\n", - "A \"skater\" is a one-line forecasting function. Yeah, I know some of you prefer twenty lines of ceremony but anyway..." + "### Step 1: Choose a skater\n", + "A \"skater\" is a one-line forecasting function. See [timemachines](https://github.com/microprediction/timemachines) README. " ], "metadata": { "id": "k27zBSRk75Pw" @@ -63,80 +63,65 @@ { "cell_type": "code", "source": [ - "from timemachines.skaters.sk.skautoarima import sk_autoarima as f\n", - "from timemachines.skaters.sk.skautoarimawiggly import sk_autoarima_wiggly_huber_d05_m3 as g" + "from timemachines.skaters.sk.skautoarima import sk_autoarima as f" ], "metadata": { "id": "oRE4vsWB74U4" }, - "execution_count": 4, + "execution_count": 1, "outputs": [] }, { "cell_type": "markdown", "source": [ - "### Step 2: Let's see if it helps make things more regular" + "## Step 2: Wiggle it\n", + "\n" ], "metadata": { - "id": "jnZSyXP18Xfg" + "id": "n4PKndHHEu5L" } }, { "cell_type": "code", "source": [ - "from timemachines.skatertools.data.ornstein import simulate_arima_like_path\n", - "from timemachines.skatertools.sensitivity.skatersensitivity import skater_bump\n", - "\n", - "\n", - "def skater_bump_plot(f, g, ndx, k):\n", - " \"\"\" Plot sensitivity to k'th to last observation,\n", - " and compare to an alternative skater g that might be smoother\n", - " \"\"\"\n", - " import numpy as np\n", - " ys = simulate_arima_like_path(seq_len=50)\n", - " y_final, x_final = skater_bump(ys=ys, f=f, ndx=ndx, k=k)\n", - " discont_max = np.max(np.diff(np.array(x_final)))\n", - " discont_median = np.median(np.abs(np.diff(np.array(x_final))))\n", - " if discont_max>5*discont_median:\n", - " print('Comparing ...')\n", - " y_alt, x_alt = skater_bump(ys=ys, f=g, ndx=ndx, k=k)\n", - " import matplotlib.pyplot as plt\n", - " plt.plot(y_final,x_final, 'rx')\n", - " plt.plot(y_alt, x_alt, 'go')\n", - " plt.ylabel('Prediction '+str(k)+' steps ahead')\n", - " kstub = g.__name__.split('_')[-1]\n", - " plt.xlabel('Value taken by y['+str(ndx)+'] w/ wiggle '+kstub)\n", - " plt.grid()\n", - " plt.title(f.__name__)\n", - " plt.legend(['original','wiggled'])\n", - " plt.show()\n", - " import time\n", + "from timemachines.skatertools.smoothing.wiggling import wiggler\n", "\n", + "def g(*args,**kwargs):\n", + " \"\"\" A more regular version of a time-series model \"\"\"\n", + " return wiggler(f=f,*args, **kwargs)" + ], + "metadata": { + "id": "a4x18PkcEw4O" + }, + "execution_count": 3, + "outputs": [] + }, + { + "cell_type": "markdown", + "source": [ + "### Step 3: Let's see if it helps make things more regular\n", + "Bootstrapping will cost you some compute. And this demo is really slow as it will compute Auto-ARIMA many times over. See you after breakfast." + ], + "metadata": { + "id": "jnZSyXP18Xfg" + } + }, + { + "cell_type": "code", + "source": [ + "from timemachines.skatertools.sensitivity.skatersensitivity import skater_bump_plot\n", "skater_bump_plot(f=f, g=g, ndx=-5, k=1)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", - "height": 636 + "height": 545 }, "id": "ktg4xxH57rNF", - "outputId": "3c3da9b2-b21c-4df0-ed72-cad575235e04" + "outputId": "2565d9b0-d0a5-452b-b579-9eb4803086c5" }, - "execution_count": 5, + "execution_count": 4, "outputs": [ - { - "metadata": { - "tags": null - }, - "name": "stderr", - "output_type": "stream", - "text": [ - "/usr/local/lib/python3.7/dist-packages/statsmodels/tsa/statespace/sarimax.py:966: UserWarning: Non-stationary starting autoregressive parameters found. Using zeros as starting parameters.\n", - " warn('Non-stationary starting autoregressive parameters'\n", - "/usr/local/lib/python3.7/dist-packages/statsmodels/base/model.py:606: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n", - " ConvergenceWarning)\n" - ] - }, { "metadata": { "tags": null @@ -144,15 +129,15 @@ "name": "stdout", "output_type": "stream", "text": [ - "{'ar': 16,\n", - " 'ar_true': 16,\n", + "{'ar': 5,\n", + " 'ar_true': 5,\n", " 'exog': 0,\n", " 'exog_variance': 0,\n", - " 'ma': 2,\n", - " 'ma_true': 2,\n", + " 'ma': 0,\n", + " 'ma_true': 0,\n", " 'measurement_variance': 0,\n", - " 'reduced_ar': 16,\n", - " 'reduced_ma': 2,\n", + " 'reduced_ar': 5,\n", + " 'reduced_ma': 0,\n", " 'seasonal_ar': 0,\n", " 'seasonal_ma': 0,\n", " 'trend': 1,\n", @@ -161,16 +146,14 @@ ] }, { - "output_type": "display_data", "data": { + "image/png": "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\n", "text/plain": [ "
" - ], - "image/png": "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\n" + ] }, - "metadata": { - "needs_background": "light" - } + "metadata": {}, + "output_type": "display_data" } ] },