{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "# Regularization: shrinkage and sparsity\n",
        "\n",
        "A companion to Lecture 14\n",
        "\n",
        "Marcel Turcotte  \n",
        "2026-10-27\n",
        "\n",
        "This notebook isolates the effect of L1 and squared L2 penalties in\n",
        "linear regression. We know the true coefficients of the synthetic data,\n",
        "so we can compare them with ordinary least squares (OLS), Ridge, and\n",
        "Lasso. OLS is the unregularized baseline; Ridge and Lasso apply L2 and\n",
        "L1 regularization, respectively. Lecture 14 applies the same penalty\n",
        "ideas to neural-network weights.\n",
        "\n",
        "The purpose is to distinguish **shrinkage** from **sparsity**. This\n",
        "experiment does not establish which estimator predicts best on unseen\n",
        "data.\n",
        "\n",
        "## Data and models\n",
        "\n",
        "Fifteen features are generated, but only three have nonzero true\n",
        "coefficients. We standardize each feature to zero sample mean and unit\n",
        "sample variance before fitting the models. The generated features\n",
        "already have comparable scales, so standardization makes only a small\n",
        "numerical difference here; it makes the preprocessing assumption\n",
        "explicit."
      ],
      "id": "86c446f5-1632-4e70-b195-2b3ddd7e07ba"
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {},
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.datasets import make_regression\n",
        "from sklearn.linear_model import LinearRegression, Ridge, Lasso\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "X, y, true_coef = make_regression(n_samples=200, n_features=15, n_informative=3,\n",
        "                                 noise=10.0, coef=True, random_state=42)\n",
        "scaler = StandardScaler()\n",
        "X_scaled = scaler.fit_transform(X)\n",
        "\n",
        "# Express the known coefficients in the same coordinates as the fitted models.\n",
        "true_coef_scaled = true_coef * scaler.scale_\n",
        "coefficient_models = {\n",
        "    'OLS': LinearRegression().fit(X_scaled, y),\n",
        "    'Ridge (L2)': Ridge(alpha=10.0).fit(X_scaled, y),\n",
        "    'Lasso (L1)': Lasso(alpha=10.0).fit(X_scaled, y)\n",
        "}"
      ],
      "id": "f58d2685"
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Why standardize?\n",
        "\n",
        "Feature units affect coefficient penalties. Multiplying a feature’s\n",
        "values by 100 allows its coefficient to be divided by 100 while\n",
        "preserving the same predictions, but changes the penalty on that\n",
        "coefficient. Standardization puts the penalties on a common scale across\n",
        "features. It does not itself create the sparsity associated with L1\n",
        "regularization. See the <a\n",
        "href=\"https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html\"\n",
        "class=\"external\" data-icon=\"true\">StandardScaler documentation</a>.\n",
        "\n",
        "For a feature transformed as $x'_j=(x_j-\\mu_j)/s_j$, its coefficient in\n",
        "standardized coordinates is $w'_j=s_jw_j$. Centering is absorbed by the\n",
        "fitted intercept. The plot and table below compare **true and estimated\n",
        "coefficients in these same standardized coordinates**: each coefficient\n",
        "describes the predicted change in the target for a\n",
        "one-standard-deviation increase in its feature, holding the other\n",
        "features fixed.\n",
        "\n",
        "All generated observations are used as training data for this\n",
        "coefficient demonstration. When evaluating prediction on held-out data,\n",
        "fit the scaler only on training data and reuse it to transform\n",
        "validation and test data; a scikit-learn pipeline keeps that\n",
        "preprocessing with the model.\n",
        "\n",
        "## Compare coefficients\n",
        "\n",
        "The informative and irrelevant features use **separate vertical\n",
        "scales**. The expanded scale in the lower panel makes small nonzero\n",
        "coefficients visible. Hollow circles identify coefficients that are\n",
        "exactly zero, measured before display rounding."
      ],
      "id": "ebf5a66a-7baf-43e9-af01-2f752c2c0e99"
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {},
      "outputs": [
        {
          "output_type": "display_data",
          "metadata": {},
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABBYAAALYCAYAAADIEHFUAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90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\nAACYRP0fC5cYhkcuzd4AAAAASUVORK5CYII=\n"
          }
        }
      ],
      "source": [
        "series = [('True', true_coef_scaled)] + [\n",
        "    (name, fitted.coef_) for name, fitted in coefficient_models.items()\n",
        "]\n",
        "colors = ['#555555', '#377eb8', '#e68613', '#319354']\n",
        "groups = [\n",
        "    (true_coef != 0, 'Informative features'),\n",
        "    (true_coef == 0, 'Irrelevant features: expanded vertical scale')\n",
        "]\n",
        "fig, axes = plt.subplots(2, 1, figsize=(11, 7.5))\n",
        "for ax, (mask, title) in zip(axes, groups):\n",
        "    feature_indices = np.flatnonzero(mask)\n",
        "    positions = np.arange(len(feature_indices))\n",
        "    for offset, ((name, values), color) in enumerate(zip(series, colors)):\n",
        "        x_positions = positions + (offset-1.5)*0.2\n",
        "        nonzero = np.count_nonzero(values)\n",
        "        ax.bar(x_positions, values[mask], width=0.2, color=color,\n",
        "               label=f'{name}: {nonzero}/15 nonzero')\n",
        "        zeros = values[mask] == 0\n",
        "        ax.scatter(x_positions[zeros], values[mask][zeros], s=35,\n",
        "                   facecolors='white', edgecolors=color, zorder=4)\n",
        "    ax.set(xticks=positions, xticklabels=feature_indices, xlabel='Feature index',\n",
        "           ylabel='Coefficient for standardized feature', title=title)\n",
        "    ax.axhline(0, color='black', linewidth=0.7)\n",
        "    ax.grid(axis='y', alpha=0.2)\n",
        "handles, labels = axes[0].get_legend_handles_labels()\n",
        "fig.legend(handles, labels, loc='upper center', ncol=2)\n",
        "fig.text(0.5, 0.01, 'Hollow circles mark exact zeros; the two panels use different vertical scales.',\n",
        "         ha='center')\n",
        "fig.tight_layout(rect=(0, 0.035, 1, 0.91))\n",
        "plt.show()"
      ],
      "id": "77eaa922"
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {},
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Feature         True        OLS      Ridge      Lasso\n",
            "0               0.00     -0.35      0.09      0.00\n",
            "1               0.00     -0.62     -0.75     -0.00\n",
            "2               0.00     -0.07      0.33      0.00\n",
            "3               0.00     -0.53     -1.06     -0.00\n",
            "4               0.00     -1.31     -1.52     -0.00\n",
            "5               0.00      0.66      0.71      0.00\n",
            "6               0.00      0.50      1.07      0.00\n",
            "7               0.00     -0.91     -0.98     -0.00\n",
            "8              23.71     24.13     23.25     15.02\n",
            "9               0.00      0.98      1.27      0.00\n",
            "10              0.00     -0.70     -1.36     -0.00\n",
            "11              0.00     -0.19     -0.55     -0.00\n",
            "12              0.00     -0.72     -1.13     -0.00\n",
            "13            106.95    105.85    100.34     95.86\n",
            "14             79.06     79.87     75.69     69.79\n",
            "\n",
            "Counts and sparsity, measured before rounding:\n",
            "Estimator      Nonzero   Exactly zero    Sparsity\n",
            "OLS                 15              0        0.0%\n",
            "Ridge (L2)          15              0        0.0%\n",
            "Lasso (L1)           3             12       80.0%\n",
            "True informative indices: [ 8 13 14]\n",
            "Lasso nonzero indices:    [ 8 13 14]"
          ]
        }
      ],
      "source": [
        "print(f\"{'Feature':<9} {'True':>10} {'OLS':>10} {'Ridge':>10} {'Lasso':>10}\")\n",
        "for index, value in enumerate(true_coef_scaled):\n",
        "    estimates = [fitted.coef_[index] for fitted in coefficient_models.values()]\n",
        "    print(f\"{index:<9} {value:>10.2f}\" + ''.join(f'{estimate:>10.2f}' for estimate in estimates))\n",
        "\n",
        "print('\\nCounts and sparsity, measured before rounding:')\n",
        "print(f\"{'Estimator':<12} {'Nonzero':>9} {'Exactly zero':>14} {'Sparsity':>11}\")\n",
        "for name, fitted in coefficient_models.items():\n",
        "    nonzero = np.count_nonzero(fitted.coef_)\n",
        "    zeros = len(fitted.coef_) - nonzero\n",
        "    sparsity = 100*zeros/len(fitted.coef_)\n",
        "    print(f'{name:<12} {nonzero:>9} {zeros:>14} {sparsity:>10.1f}%')\n",
        "print('True informative indices:', np.flatnonzero(true_coef))\n",
        "print('Lasso nonzero indices:   ', np.flatnonzero(coefficient_models['Lasso (L1)'].coef_))"
      ],
      "id": "66894937"
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "With these fixed settings, OLS and Ridge retain fifteen nonzero\n",
        "coefficients and Lasso retains three, giving 80% exact-zero sparsity.\n",
        "Here Lasso selects the three informative features. This is an\n",
        "illustration, not a guarantee of recovering true features in other\n",
        "datasets, particularly when predictors are correlated.\n",
        "\n",
        "L2 encourages smaller weights collectively and generally retains a dense\n",
        "solution. It penalizes the total squared coefficient magnitude; an\n",
        "individual coefficient need not become smaller than its OLS estimate\n",
        "when predictors are correlated. Even independently generated features\n",
        "have some correlation in a finite sample.\n",
        "\n",
        "L1 encourages sparsity: the corner of the absolute-value penalty at zero\n",
        "allows the optimum to have exactly zero coefficients, and Lasso’s\n",
        "coordinate-descent solver can find these sparse solutions. These zeros\n",
        "are a property of the optimization problem. Display rounding is not used\n",
        "to decide whether a coefficient is zero. See the <a\n",
        "href=\"https://scikit-learn.org/stable/modules/linear_model.html#lasso\"\n",
        "class=\"external\" data-icon=\"true\">Lasso user guide</a>.\n",
        "\n",
        "## What do the penalty coefficients mean?\n",
        "\n",
        "The two `alpha=10` values were chosen for a qualitative demonstration.\n",
        "They are not matched penalty strengths. <a\n",
        "href=\"https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Ridge.html\"\n",
        "class=\"external\" data-icon=\"true\">Ridge</a> minimizes\n",
        "\n",
        "$$\n",
        "\\|y-Xw\\|_2^2+\\alpha\\|w\\|_2^2,\n",
        "$$\n",
        "\n",
        "whereas <a\n",
        "href=\"https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Lasso.html\"\n",
        "class=\"external\" data-icon=\"true\">Lasso</a> minimizes\n",
        "\n",
        "$$\n",
        "\\frac{1}{2N}\\|y-Xw\\|_2^2+\\alpha\\|w\\|_1.\n",
        "$$\n",
        "\n",
        "Here $X$ denotes the standardized feature matrix and $w$ its\n",
        "coefficients. The fitted intercept is omitted from these compact\n",
        "expressions and is not penalized. Different data-loss normalizations and\n",
        "penalty definitions mean that coefficients from different APIs need\n",
        "interpretation before comparison.\n",
        "\n",
        "## Connection to neural networks\n",
        "\n",
        "The penalty idea extends to entries of the neural network’s weight\n",
        "matrices. In `SimpleMLP`, the objective is mean BCE plus $\\lambda/2$\n",
        "times the sum of squared weights, and the added gradient is $\\lambda W$.\n",
        "Biases are not penalized.\n",
        "\n",
        "L1 encourages sparse connections in a neural network, but ordinary\n",
        "gradient updates need not return exact zeros. An input feature is fully\n",
        "disconnected only when all of its outgoing weights are zero; a single\n",
        "zero connection does not establish feature removal.\n",
        "\n",
        "## What this experiment establishes\n",
        "\n",
        "The figures illustrate coefficient shrinkage and sparsity. The three\n",
        "estimators retain different numbers of features; similar-looking\n",
        "coefficients on the informative features do not establish equal\n",
        "predictive performance. Nor does a sparser model automatically predict\n",
        "better.\n",
        "\n",
        "To compare prediction on unseen examples, use separate training,\n",
        "validation, and test data, choose each estimator’s penalty using\n",
        "validation results, and report final test errors. Regularization that is\n",
        "too strong can underfit. The fixed penalty values here serve the\n",
        "coefficient demonstration and have not been selected for prediction\n",
        "accuracy.\n",
        "\n",
        "[Return to Lecture 14](index.qmd)."
      ],
      "id": "b1f98ad8-5382-4e4f-9961-3474170a4fa7"
    }
  ],
  "nbformat": 4,
  "nbformat_minor": 5,
  "metadata": {
    "kernelspec": {
      "name": "csi4106-pytorch",
      "display_name": "CSI4106 (PyTorch)",
      "language": "python",
      "path": "/Users/turcotte/Library/Jupyter/kernels/csi4106-pytorch"
    },
    "language_info": {
      "name": "python",
      "codemirror_mode": {
        "name": "ipython",
        "version": "3"
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.12.12"
    }
  }
}