zgulde/regression_recap.ipynb

## regression_recap.ipynb
{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "41e5840c-ecab-404e-af46-6e6b1067307f",
   "metadata": {},
   "source": [
    "# Modeling\n",
    "\n",
    "Regression Models:\n",
    "\n",
    "- Baseline\n",
    "- LinearRegression (OLS)\n",
    "- LassoLars\n",
    "- GLM: generalized linear models -- fancy math and stats- GLM: generalized linear models -- fancy math and stats; if the relationship between target and predictors (i.e. features, or independent variables) is non-linear\n",
    "- PolynomialFeatures*: non-linear relationship between predictors and target\n",
    "    - creates interaction terms and squared terms (or even higher degrees)\n",
    "    - linear model == y ~ x1 + x2 + x3\n",
    "    - polynomial == y ~ x1^2 + x1*x2 + x2^2 + x2*x3 + x3^2 + x1 + x2 + x3\n",
    "    - total_charges ~ tenure + monthly_charges\n",
    "    - total_charges ~ tenure*monthly_charges\n",
    "\n",
    "What do we do in the modeling stage of the pipeline?\n",
    "\n",
    "1. Create a model\n",
    "    - research hyperparms\n",
    "    - try out lots of diff hyperparms\n",
    "    - baseline\n",
    "        - mean\n",
    "        - median\n",
    "1. Evaluate the model's performance\n",
    "    - visualize residuals\n",
    "    - visualize the distribution of model predictions vs actual values\n",
    "    - fancy dataframes to hold on to predictions\n",
    "1. Repeat\n",
    "1. Compare model performance\n",
    "    - fancy dataframes to hold on to predictions\n",
    "    - seperate dataframe to hold regression evaluation metrics\n",
    "    - visualize different model residuals against each other\n",
    "1. Evaluate the best model on test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "bdb1fdf6-b2e0-4ea1-9f8b-8d1577debf69",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import wrangle\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from sklearn.linear_model import LinearRegression, LassoLars\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1912421b-62f9-4979-91f1-5d6fd0156793",
   "metadata": {},
   "outputs": [],
   "source": [
    "path = 'https://gist.githubusercontent.com/ryanorsinger/55ccfd2f7820af169baea5aad3a9c60d/raw/da6c5a33307ed7ee207bd119d3361062a1d1c07e/student-mat.csv'\n",
    "\n",
    "(\n",
    "    df, X_train_exp, X_train, y_train,\n",
    "    X_validate, y_validate, X_test, y_test,\n",
    ") = wrangle.wrangle_student_math(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4a57e8ae-0f87-4a78-bdd5-30cc70cb6b7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "baseline = y_train.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0949999c-2b53-4700-afa6-1672b920f99b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Turn y_train into a dataframe so that we can store, \"hold on to\", our models' predictions\n",
    "\n",
    "y_train = pd.DataFrame({'actual': y_train})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1e2e688e-97ad-4b25-9783-f65c3cd15cd2",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train['baseline'] = baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "23432fa6-0140-41d1-a26e-e6b922e61cb9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train RMSE Baseline model:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4.498925523895268"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Train RMSE Baseline model:')\n",
    "math.sqrt(mean_squared_error(y_train.actual, y_train.baseline))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7c7121d0-c974-464b-a539-631531b5b58c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# OLS -- ordinary least squares - way the coefficients are calculated\n",
    "model2 = LinearRegression()\n",
    "# When add additional values to y_train, we need to fit with a single column\n",
    "model2.fit(X_train, y_train.actual)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "60f6c92a-3cdf-43f6-ada0-b35fd685b33d",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train['model2'] = model2.predict(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "60751519-58ef-43ed-adb7-4b8c534f6421",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train RMSE for model2:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1.7503546500121143"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Train RMSE for model2:')\n",
    "math.sqrt(mean_squared_error(y_train.actual, y_train.model2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c3e6b99c-1d2e-4b4f-b2cb-665e7a2a83ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LassoLars(alpha=0.1)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model3 = LassoLars(alpha=.1)\n",
    "model3.fit(X_train, y_train.actual)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "adbb11ca-520e-4628-b1d2-1640e8ff2d44",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train['model3'] = model3.predict(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "b35191b7-8ec0-4eac-ade6-d0ebc2a04718",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train RMSE for model3:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "2.396115747537386"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Train RMSE for model3:')\n",
    "math.sqrt(mean_squared_error(y_train.actual, y_train.model3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "32228808-6944-4f7f-9638-8d441bd11f28",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_validate = pd.DataFrame({'actual': y_validate})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "e3597b3a-8053-4237-a8af-44ecb5f987f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>actual</th>\n",
       "      <th>baseline</th>\n",
       "      <th>model2</th>\n",
       "      <th>model3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>241</th>\n",
       "      <td>12</td>\n",
       "      <td>10.524887</td>\n",
       "      <td>10.219690</td>\n",
       "      <td>10.607736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>235</th>\n",
       "      <td>10</td>\n",
       "      <td>10.524887</td>\n",
       "      <td>9.423179</td>\n",
       "      <td>9.199304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>369</th>\n",
       "      <td>11</td>\n",
       "      <td>10.524887</td>\n",
       "      <td>12.013069</td>\n",
       "      <td>11.311952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>217</th>\n",
       "      <td>8</td>\n",
       "      <td>10.524887</td>\n",
       "      <td>4.643292</td>\n",
       "      <td>7.086656</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>331</th>\n",
       "      <td>14</td>\n",
       "      <td>10.524887</td>\n",
       "      <td>13.636035</td>\n",
       "      <td>12.720384</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     actual   baseline     model2     model3\n",
       "241      12  10.524887  10.219690  10.607736\n",
       "235      10  10.524887   9.423179   9.199304\n",
       "369      11  10.524887  12.013069  11.311952\n",
       "217       8  10.524887   4.643292   7.086656\n",
       "331      14  10.524887  13.636035  12.720384"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_validate['baseline'] = y_train.actual.mean()\n",
    "y_validate['model2'] = model2.predict(X_validate)\n",
    "y_validate['model3'] = model3.predict(X_validate)\n",
    "\n",
    "y_validate.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "7713bff9-8e48-45df-bb02-1f6940e069ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train RMSE baseline:\n",
      "4.498925523895268\n",
      "Train RMSE model2:\n",
      "1.7503546500121143\n",
      "Train RMSE model3:\n",
      "2.396115747537386\n"
     ]
    }
   ],
   "source": [
    "print('Train RMSE baseline:')\n",
    "print(math.sqrt(mean_squared_error(y_train.actual, y_train.baseline)))\n",
    "\n",
    "print('Train RMSE model2:')\n",
    "print(math.sqrt(mean_squared_error(y_train.actual, y_train.model2)))\n",
    "\n",
    "print('Train RMSE model3:')\n",
    "print(math.sqrt(mean_squared_error(y_train.actual, y_train.model3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "cc3c4807-562e-461a-9a87-04bf6ba4eb57",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Validate RMSE baseline:\n",
      "4.578916932633144\n",
      "Validate RMSE model2:\n",
      "2.1264081323553436\n",
      "Validate RMSE model3:\n",
      "2.486601997892303\n"
     ]
    }
   ],
   "source": [
    "print('Validate RMSE baseline:')\n",
    "print(math.sqrt(mean_squared_error(y_validate.actual, y_validate.baseline)))\n",
    "\n",
    "print('Validate RMSE model2:')\n",
    "print(math.sqrt(mean_squared_error(y_validate.actual, y_validate.model2)))\n",
    "\n",
    "print('Validate RMSE model3:')\n",
    "print(math.sqrt(mean_squared_error(y_validate.actual, y_validate.model3)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e610cdb9-f161-426d-a8be-7e58de5b9bde",
   "metadata": {},
   "source": [
    "Looks like model 2 is the winner!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "404209a5-d67d-4fe2-9da4-3d383f05b62b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.9229879901741005"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "math.sqrt(mean_squared_error(model2.predict(X_test), y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92048e1b-6d1d-49c3-9791-d5c35b955997",
   "metadata": {},
   "source": [
    "^ this number tells us how we would expect our model to perform on future data, on future predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "94afb024-33d6-47f9-86e7-0b1b9af0d55a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# data for a single student\n",
    "unknown_student = X_test.iloc[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "a5fb59e6-b2ee-44d1-9b53-9949d8cc4192",
   "metadata": {},
   "outputs": [],
   "source": [
    "# the data needs to be in the shape that the model expects, i.e.\n",
    "# it needs to have the same number of columns as the training X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "7211aefa-13c5-4a33-81ba-37d0e81d2867",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rows, cols -- -1 means fill the extra space\n",
    "student_X = unknown_student.values.reshape(1, -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "333434bc-d827-4e65-92dd-92a982544b71",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([7.4726451])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.predict(student_X)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef2eb1d9-8f5a-4d66-9de8-5b416aeafc65",
   "metadata": {},
   "source": [
    "-----------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "ca48eda4-1e2a-42a0-aff1-2941935bdfb9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['1', 'age', 'G1', 'G2', 'age^2', 'age G1', 'age G2', 'G1^2', 'G1 G2', 'G2^2']"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# different from our other models in that this just transforms the X\n",
    "# *after* the polynomial transformation, run the transformed data through the same model(s)\n",
    "X_train_poly = X_train[['age', 'G1', 'G2']]\n",
    "\n",
    "poly = PolynomialFeatures()\n",
    "poly.fit(X_train_poly)\n",
    "X_transformed = poly.transform(X_train_poly)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "e8768085-6955-4fb9-b049-3a35407b9619",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>age</th>\n",
       "      <th>G1</th>\n",
       "      <th>G2</th>\n",
       "      <th>age^2</th>\n",
       "      <th>age G1</th>\n",
       "      <th>age G2</th>\n",
       "      <th>G1^2</th>\n",
       "      <th>G1 G2</th>\n",
       "      <th>G2^2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.357143</td>\n",
       "      <td>0.578947</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.127551</td>\n",
       "      <td>0.206767</td>\n",
       "      <td>0.335180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.714286</td>\n",
       "      <td>0.789474</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.263158</td>\n",
       "      <td>0.510204</td>\n",
       "      <td>0.563910</td>\n",
       "      <td>0.623269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0.027778</td>\n",
       "      <td>0.083333</td>\n",
       "      <td>0.087719</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.263158</td>\n",
       "      <td>0.277008</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     1       age        G1        G2     age^2    age G1    age G2      G1^2  \\\n",
       "0  1.0  0.000000  0.357143  0.578947  0.000000  0.000000  0.000000  0.127551   \n",
       "1  1.0  0.333333  0.714286  0.789474  0.111111  0.238095  0.263158  0.510204   \n",
       "2  1.0  0.166667  0.500000  0.526316  0.027778  0.083333  0.087719  0.250000   \n",
       "\n",
       "      G1 G2      G2^2  \n",
       "0  0.206767  0.335180  \n",
       "1  0.563910  0.623269  \n",
       "2  0.263158  0.277008  "
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(X_transformed, columns=poly.get_feature_names(input_features=X_train_poly.columns)).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "8a074fc0-7846-4525-bce2-3b33cc4bd870",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model4 = LinearRegression()\n",
    "model4.fit(X_transformed, y_train.actual)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "a79e6d55-d233-40a3-8b73-a32ef38ffc90",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.8184368425448687"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train['model4'] = model4.predict(X_transformed)\n",
    "\n",
    "math.sqrt(mean_squared_error(y_train.actual, y_train.model4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "7814dfe0-6b11-43ca-88db-763685e12b8a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# when we are making predictions with a model, any predictors (i.e. new data)\n",
    "# including polynomial transformation\n",
    "\n",
    "X_validate_poly = X_validate[['age', 'G1', 'G2']]\n",
    "X_validate_transformed = poly.transform(X_validate_poly)\n",
    "y_validate['model4'] = model4.predict(X_validate_transformed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "2b66b341-76e0-4945-b21d-98912e5c5aa7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2.231228440521658"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "math.sqrt(mean_squared_error(y_validate.actual, y_validate.model4))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b06d8a55-2946-4061-9bc7-1d9f72859b2c",
   "metadata": {},
   "source": [
    "Visualize Residuals\n",
    "\n",
    "residuals = actual - predicted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "8446adf5-7d2e-47de-9414-4045a6f7308b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='Frequency'>"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_train.actual.plot.hist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "14c9482b-5319-41d3-a5b0-ad1481d95886",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.scatter(y_train.actual, y_train.actual - y_train.baseline, label='baseline')\n",
    "plt.figure(figsize=(13, 7))\n",
    "plt.scatter(y_train.actual, y_train.actual - y_train.model2, label='model2 (LinearRegression)', alpha=.6)\n",
    "plt.scatter(y_train.actual, y_train.actual - y_train.model4, label='model4 (Polynomial)', alpha=.6)\n",
    "plt.hlines(0, 0, 20, ls=':', label='perfect prediction', color='black')\n",
    "plt.ylabel('Residual')\n",
    "plt.xlabel('Actual')\n",
    "plt.legend()\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "bd2a7211-8765-422e-9dbd-7e6d6d3518b8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.scatter(y_train.actual, y_train.actual - y_train.baseline, label='baseline')\n",
    "plt.figure(figsize=(13, 7))\n",
    "plt.scatter(y_train.actual, y_train.model2, label='model2 (LinearRegression)', alpha=.6)\n",
    "plt.scatter(y_train.actual, y_train.model4, label='model4 (Polynomial)', alpha=.6)\n",
    "plt.ylabel('Predicted Value')\n",
    "plt.plot([0, 20], [0, 20], ls=':', c='black', label='perfect prediction')\n",
    "plt.xlabel('Actual')\n",
    "plt.legend()\n",
    "plt.grid()"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}