{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d02d91ee",
   "metadata": {},
   "source": [
    "# Assignment 2: worked solution\n",
    "\n",
    "This notebook implements simple linear regression from scratch using one feature.\n",
    "\n",
    "The goal is not to memorize the code. Follow how training turns many rows into two learned\n",
    "values—the intercept and slope—and how prediction uses those values.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4ce65d51",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train 256  test 86\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "CSV = \"../data/penguins.csv\"\n",
    "penguins = pd.read_csv(CSV).dropna(\n",
    "    subset=[\"flipper_length_mm\", \"body_mass_g\"]\n",
    ")\n",
    "\n",
    "Xr = penguins[[\"flipper_length_mm\"]]\n",
    "yr = penguins[\"body_mass_g\"]\n",
    "\n",
    "Xr_tr, Xr_te, yr_tr, yr_te = train_test_split(\n",
    "    Xr, yr, test_size=0.25, random_state=42\n",
    ")\n",
    "\n",
    "print(\"train\", len(Xr_tr), \" test\", len(Xr_te))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e232f96d",
   "metadata": {},
   "source": [
    "## Task 1: `fit_simple_linear_regression`\n",
    "\n",
    "Training calculates two values from the training rows:\n",
    "\n",
    "- the slope describes how predicted body mass changes with flipper length;\n",
    "- the intercept positions the fitted line.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a97af2df",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fit_simple_linear_regression(x, y):\n",
    "    \"\"\"Return the least-squares intercept and slope for one feature.\"\"\"\n",
    "    x = np.asarray(x, dtype=float).reshape(-1)\n",
    "    y = np.asarray(y, dtype=float).reshape(-1)\n",
    "\n",
    "    x_mean = x.mean()\n",
    "    y_mean = y.mean()\n",
    "\n",
    "    slope_numerator = np.sum((x - x_mean) * (y - y_mean))\n",
    "    slope_denominator = np.sum((x - x_mean) ** 2)\n",
    "\n",
    "    # if slope_denominator == 0:\n",
    "    #     raise ValueError(\"x must contain at least two different values\")\n",
    "\n",
    "    slope = slope_numerator / slope_denominator\n",
    "    intercept = y_mean - slope * x_mean\n",
    "    return intercept, slope\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e29fff9",
   "metadata": {},
   "source": [
    "## Task 2: `predict_simple_linear_regression`\n",
    "\n",
    "After training, prediction is short: substitute each new flipper length into the fitted line.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "36d65fcd",
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_simple_linear_regression(x, intercept, slope):\n",
    "    \"\"\"Predict y values using a fitted intercept and slope.\"\"\"\n",
    "    x = np.asarray(x, dtype=float).reshape(-1)\n",
    "    return intercept + slope * x\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f66d8875",
   "metadata": {},
   "source": [
    "## Task 3: `mean_absolute_error_from_scratch`\n",
    "\n",
    "MAE is the average size of the prediction errors, ignoring whether each error is positive or\n",
    "negative.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "8fc2e26f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def mean_absolute_error_from_scratch(y_true, y_pred):\n",
    "    \"\"\"Return the average absolute prediction error.\"\"\"\n",
    "    y_true = np.asarray(y_true, dtype=float)\n",
    "    y_pred = np.asarray(y_pred, dtype=float)\n",
    "    return np.mean(np.abs(y_true - y_pred))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c08af91",
   "metadata": {},
   "source": [
    "## Check the solution\n",
    "\n",
    "The small example has a known answer. After that check passes, fit the model using the penguin\n",
    "training rows and evaluate it on the unseen test rows.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c0bf3e6a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "OK  fit -> intercept 1, slope 2\n",
      "OK  predict -> [ 9. 11.]\n",
      "OK  MAE -> 0.5\n",
      "penguin intercept: -5741.26\n",
      "penguin slope:     49.506 grams per mm\n",
      "penguin test MAE:  282.0 g\n"
     ]
    }
   ],
   "source": [
    "toy_x = np.array([0, 1, 2, 3], dtype=float)\n",
    "toy_y = np.array([1, 3, 5, 7], dtype=float)\n",
    "\n",
    "toy_intercept, toy_slope = fit_simple_linear_regression(toy_x, toy_y)\n",
    "assert np.isclose(toy_intercept, 1.0)\n",
    "assert np.isclose(toy_slope, 2.0)\n",
    "print(\"OK  fit -> intercept 1, slope 2\")\n",
    "\n",
    "toy_prediction = predict_simple_linear_regression(\n",
    "    np.array([4, 5]), toy_intercept, toy_slope\n",
    ")\n",
    "assert np.allclose(toy_prediction, [9, 11])\n",
    "print(\"OK  predict ->\", toy_prediction)\n",
    "\n",
    "toy_mae = mean_absolute_error_from_scratch([1, 2], [1.5, 1.5])\n",
    "assert np.isclose(toy_mae, 0.5)\n",
    "print(\"OK  MAE ->\", toy_mae)\n",
    "\n",
    "scratch_intercept, scratch_slope = fit_simple_linear_regression(\n",
    "    Xr_tr[\"flipper_length_mm\"], yr_tr\n",
    ")\n",
    "scratch_prediction = predict_simple_linear_regression(\n",
    "    Xr_te[\"flipper_length_mm\"], scratch_intercept, scratch_slope\n",
    ")\n",
    "scratch_mae = mean_absolute_error_from_scratch(yr_te, scratch_prediction)\n",
    "\n",
    "print(f\"penguin intercept: {scratch_intercept:.2f}\")\n",
    "print(f\"penguin slope:     {scratch_slope:.3f} grams per mm\")\n",
    "print(f\"penguin test MAE:  {scratch_mae:.1f} g\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bbe8f751",
   "metadata": {},
   "source": [
    "## Visual check\n",
    "\n",
    "A fitted line should pass through the middle of the training points. The test points were not\n",
    "used to calculate the line.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a0143da4",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 750x450 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(7.5, 4.5))\n",
    "\n",
    "ax.scatter(\n",
    "    Xr_tr[\"flipper_length_mm\"],\n",
    "    yr_tr,\n",
    "    alpha=0.45,\n",
    "    s=22,\n",
    "    label=\"training penguins\",\n",
    ")\n",
    "ax.scatter(\n",
    "    Xr_te[\"flipper_length_mm\"],\n",
    "    yr_te,\n",
    "    alpha=0.8,\n",
    "    s=24,\n",
    "    color=\"tab:orange\",\n",
    "    label=\"test penguins\",\n",
    ")\n",
    "\n",
    "line_x = np.linspace(\n",
    "    Xr[\"flipper_length_mm\"].min(),\n",
    "    Xr[\"flipper_length_mm\"].max(),\n",
    "    100,\n",
    ")\n",
    "line_y = predict_simple_linear_regression(\n",
    "    line_x, scratch_intercept, scratch_slope\n",
    ")\n",
    "ax.plot(line_x, line_y, color=\"tab:red\", linewidth=2.5, label=\"fitted line\")\n",
    "\n",
    "ax.set_xlabel(\"flipper length (mm)\")\n",
    "ax.set_ylabel(\"body mass (g)\")\n",
    "ax.set_title(\"From-scratch linear regression\")\n",
    "ax.legend(frameon=False)\n",
    "ax.grid(alpha=0.2)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b7a2559",
   "metadata": {},
   "source": [
    "## Main ideas\n",
    "\n",
    "- kNN stores training examples and does most of its calculation during prediction.\n",
    "- Linear regression calculates an intercept and slope during training.\n",
    "- A simple fitted regression model stores two numbers instead of all training rows.\n",
    "- Test rows are used for evaluation, not for calculating the fitted line.\n",
    "- A small example with a known answer is useful for checking a from-scratch implementation.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "79ad06cb-979b-4fbc-98aa-730ed07e2420",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ML Workshop (Python 3.11)",
   "language": "python",
   "name": "mlworkshop"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}
