{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Random Search\n", "> In this chapter you will be introduced to another popular automated hyperparameter tuning methodology called Random Search. You will learn what it is, how it works and importantly how it differs from grid search. You will learn some advantages and disadvantages of this method and when to choose this method compared to Grid Search. You will practice undertaking a Random Search with Scikit Learn as well as visualizing & interpreting the output. This is the Summary of lecture \"Hyperparameter Tuning in Python\", via datacamp.\n", "\n", "- toc: true \n", "- badges: true\n", "- comments: true\n", "- author: Chanseok Kang\n", "- categories: [Python, Datacamp, Machine_Learning]\n", "- image: images/grid_vs_random.png" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Introducting Random Search\n", "- Similar to grid search:\n", " - Define an estimator, which hyperparameters to tune and the range of values for each hyperparameter.\n", " - Set a Cross-Validation scheme and scoring function\n", "> Note - This paper shows empirically and theoretically that randomly chosen trials are more efficient for hyperparmeter optimization than trials on a grid search (Bengio & Bergstra (2012))\n", "- Two main reason:\n", " 1. Not every hyperparameter is as important\n", " 2. A little trick of probability" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Randomly Sample Hyperparameters\n", "To undertake a random search, we firstly need to undertake a random sampling of our hyperparameter space.\n", "\n", "In this exercise, you will firstly create some lists of hyperparameters that can be zipped up to a list of lists. Then you will randomly sample hyperparameter combinations preparation for running a random search.\n", "\n", "You will use just the hyperparameters `learning_rate` and `min_samples_leaf` of the GBM algorithm to keep the example illustrative and not overly complicated." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[[1.455075376884422, 25],\n", " [0.9833668341708542, 17],\n", " [1.2678894472361808, 21],\n", " [1.2604020100502513, 31],\n", " [0.48170854271356783, 23],\n", " [1.2304522613065327, 25],\n", " [0.4966834170854271, 26],\n", " [0.1672361809045226, 13],\n", " [1.455075376884422, 33],\n", " [0.6614070351758794, 30],\n", " [1.050753768844221, 33],\n", " [0.07738693467336683, 38],\n", " [0.06241206030150754, 19],\n", " [0.818643216080402, 12],\n", " [0.4592462311557789, 23],\n", " [1.0432663316582915, 20],\n", " [0.5266331658291458, 18],\n", " [0.8860301507537688, 34],\n", " [0.5940201005025125, 24],\n", " [0.2570854271356784, 25],\n", " [0.8411055276381909, 22],\n", " [0.5490954773869346, 31],\n", " [1.2604020100502513, 21],\n", " [0.7886934673366834, 19],\n", " [0.7737185929648241, 36],\n", " [1.4401005025125628, 21],\n", " [0.9908542713567839, 35],\n", " [1.222964824120603, 34],\n", " [1.2529145728643216, 30],\n", " [0.17472361809045225, 32]]\n" ] } ], "source": [ "from itertools import product\n", "from pprint import pprint\n", "\n", "# Create a list of values for the learning_rate hyperparameter\n", "learn_rate_list = list(np.linspace(0.01, 1.5, 200))\n", "\n", "# Create a list of values for the min_samples_leaf hyperparameter\n", "min_samples_list = list(range(10, 41))\n", "\n", "# Combination list\n", "combinations_list = [list(x) for x in product(learn_rate_list, min_samples_list)]\n", "\n", "# Sample hyperparameter combinations for a randomsearch\n", "random_combinations_index = np.random.choice(range(0, len(combinations_list)), \n", " 30, replace=False)\n", "combinations_random_chosen = [combinations_list[x] for x in random_combinations_index]\n", "\n", "# Print the result\n", "pprint(combinations_random_chosen)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Randomly Search with Random Forest\n", "To solidify your knowledge of random sampling, let's try a similar exercise but using different hyperparameters and a different algorithm.\n", "\n", "As before, create some lists of hyperparameters that can be zipped up to a list of lists. You will use the hyperparameters `criterion`, `max_depth` and `max_features` of the random forest algorithm. Then you will randomly sample hyperparameter combinations in preparation for running a random search." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[['entropy', 'auto', 26],\n", " ['gini', None, 26],\n", " ['entropy', 'sqrt', 9],\n", " ['gini', 'log2', 28],\n", " ['entropy', 'auto', 54],\n", " ['entropy', None, 39],\n", " ['entropy', 'sqrt', 5],\n", " ['gini', 'sqrt', 51],\n", " ['gini', None, 34],\n", " ['gini', 'auto', 54],\n", " ['entropy', 'log2', 3],\n", " ['entropy', None, 29],\n", " ['gini', 'auto', 25],\n", " ['gini', 'sqrt', 9],\n", " ['gini', 'log2', 6],\n", " ['entropy', 'log2', 52],\n", " ['entropy', None, 22],\n", " ['entropy', 'sqrt', 32],\n", " ['entropy', 'auto', 3],\n", " ['gini', 'auto', 41],\n", " ['entropy', 'sqrt', 39],\n", " ['gini', 'sqrt', 43],\n", " ['gini', 'sqrt', 41],\n", " ['gini', None, 23],\n", " ['gini', 'auto', 11],\n", " ['gini', 'auto', 52],\n", " ['gini', 'auto', 49],\n", " ['entropy', 'log2', 25],\n", " ['gini', None, 31],\n", " ['gini', None, 27]]\n" ] } ], "source": [ "import random\n", "\n", "# Create lists for criterion and max_features\n", "criterion_list = ['gini', 'entropy']\n", "max_feature_list = ['auto', 'sqrt', 'log2', None]\n", "\n", "# Create a list of values for the max_dep hyperparameter\n", "max_depth_list = list(range(3, 56))\n", "\n", "# Combination list\n", "combinations_list = [list(x) for x in product(criterion_list, max_feature_list, max_depth_list)]\n", "\n", "# Cample hyperparameter combinations for a random search\n", "combinations_random_chosen = random.sample(combinations_list, 30)\n", "\n", "# Print the result\n", "pprint(combinations_random_chosen)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualizing a Random Search\n", "Visualizing the search space of random search allows you to easily see the coverage of this technique and therefore allows you to see the effect of your sampling on the search space.\n", "\n", "In this exercise you will use several different samples of hyperparameter combinations and produce visualizations of the search space.\n", "\n", "The function `sample_and_visualize_hyperparameters()` takes a single argument (number of combinations to sample) and then randomly samples hyperparameter combinations, just like you did in the last exercise! The function will then visualize the combinations." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "\n", "def sample_and_visualize_hyperparameters(n_samples):\n", " # If asking for all combinations, just return the entire list.\n", " if n_samples == len(combinations_list):\n", " combinations_random_chosen = combinations_list\n", " else:\n", " combinations_random_chosen = []\n", " random_combinations_index = np.random.choice(range(0, len(combinations_list)), n_samples, replace=False)\n", " combinations_random_chosen = [combinations_list[x] for x in random_combinations_index]\n", " \n", " # Pull out the X and Y to plot\n", " rand_y, rand_x = [x[0] for x in combinations_random_chosen], [x[1] for x in combinations_random_chosen]\n", "\n", " # Plot \n", " plt.clf() \n", " plt.scatter(rand_y, rand_x, c=['blue']*len(combinations_random_chosen))\n", " plt.gca().set(xlabel='learn_rate', ylabel='min_samples_leaf', title='Random Search Hyperparameters')\n", " plt.gca().set_xlim([0.01, 1.5])\n", " plt.gca().set_ylim([10, 29])" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "4000\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Create a list of values for the learning_rate hyperparameter\n", "learn_rate_list = list(np.linspace(0.01, 1.5, 200))\n", "\n", "# Create a list of values for the min_samples_leaf hyperparameter\n", "min_samples_list = list(range(10, 30))\n", "\n", "# Combination list\n", "combinations_list = [list(x) for x in product(learn_rate_list, min_samples_list)]\n", "\n", "# Confirm how many hyperparameter combinations & print\n", "number_combs = len(combinations_list)\n", "print(number_combs)\n", "\n", "# Sample and visualize specified combinations\n", "sample_and_visualize_hyperparameters(50)" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Sample and visualize specified combinations\n", "sample_and_visualize_hyperparameters(500)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Sample and visualize specified combinations\n", "sample_and_visualize_hyperparameters(1500)" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Sample all the hyperparameter combinations & visualize\n", "sample_and_visualize_hyperparameters(number_combs)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Random Search in Scikit Learn\n", "- Comparing to GridSearchCV\n", " 1. Decide an algorithm/estimator\n", " 2. Define which hyperparameters we will tune\n", " 3. Define a range of values for each hyperparameter\n", " 4. Setting a Cross-Validation scheme\n", " 5. Define a score function\n", " 6. Include extra useful information or function\n", "- In Random Search,\n", " 7. Decide how many samples to take and sample it" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### The RandomizedSearchCV Object\n", "Just like the `GridSearchCV` library from Scikit Learn, `RandomizedSearchCV` provides many useful features to assist with efficiently undertaking a random search. You're going to create a `RandomizedSearchCV` object, making the small adjustment needed from the `GridSearchCV` object.\n", "\n", "The desired options are:\n", "\n", "- A default Gradient Boosting Classifier Estimator\n", "- 5-fold cross validation\n", "- Use accuracy to score the models\n", "- Use 4 cores for processing in parallel\n", "- Ensure you refit the best model and return training scores\n", "- Randomly sample 10 models\n", "\n", "The hyperparameter grid should be for `learning_rate` (150 values between 0.1 and 2) and `min_samples_leaf` (all values between and including 20 and 64)." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "from sklearn.model_selection import train_test_split\n", "\n", "credit_card = pd.read_csv('./dataset/credit-card-full.csv')\n", "# To change categorical variable with dummy variables\n", "credit_card = pd.get_dummies(credit_card, columns=['SEX', 'EDUCATION', 'MARRIAGE'], drop_first=True)\n", "\n", "X = credit_card.drop(['ID', 'default payment next month'], axis=1)\n", "y = credit_card['default payment next month']\n", "\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, shuffle=True)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[0.5845637583892618 1.5536912751677854 0.6483221476510067\n", " 1.0563758389261746 1.8469798657718122 1.2348993288590604 0.1\n", " 1.3114093959731543 0.41879194630872485 0.3550335570469798]\n", "[51 54 36 45 33 45 26 59 38 23]\n" ] } ], "source": [ "from sklearn.ensemble import GradientBoostingClassifier\n", "from sklearn.model_selection import RandomizedSearchCV\n", "\n", "# Create the parameter grid\n", "param_grid = {'learning_rate': np.linspace(0.1, 2, 150), \n", " 'min_samples_leaf': list(range(20, 65))}\n", "\n", "# Create a random search object\n", "random_GBM_class = RandomizedSearchCV(\n", " estimator=GradientBoostingClassifier(),\n", " param_distributions=param_grid,\n", " n_iter=10,\n", " scoring='accuracy', n_jobs=4, cv=5, refit=True, return_train_score=True\n", ")\n", "\n", "# Fit to the training data\n", "random_GBM_class.fit(X_train, y_train)\n", "\n", "# Print the values used for both hyperparameters\n", "print(random_GBM_class.cv_results_['param_learning_rate'])\n", "print(random_GBM_class.cv_results_['param_min_samples_leaf'])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### RandomizedSearchCV in Scikit Learn\n", "Let's practice building a `RandomizedSearchCV` object using Scikit Learn.\n", "\n", "The hyperparameter grid should be for `max_depth` (all values between and including 5 and 25) and `max_features` ('auto' and 'sqrt').\n", "\n", "The desired options for the `RandomizedSearchCV` object are:\n", "\n", "- A RandomForestClassifier Estimator with `n_estimators` of 80.\n", "- 3-fold cross validation (`cv`)\n", "- Use `roc_auc` to score the models\n", "- Use 4 cores for processing in parallel (`n_jobs`)\n", "- Ensure you refit the best model and return training scores\n", "- Only sample 5 models for efficiency (`n_iter`)\n", "\n", "Remember, to extract the chosen hyperparameters these are found in `cv_results_` with a column per hyperparameter. For example, the column for the hyperparameter `criterion` would be `param_criterion`." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[5 19 15 14 15]\n", "['sqrt' 'auto' 'auto' 'auto' 'sqrt']\n" ] } ], "source": [ "from sklearn.ensemble import RandomForestClassifier\n", "\n", "# Create the parameter grid\n", "param_grid = {'max_depth': list(range(5, 26)), 'max_features': ['auto', 'sqrt']}\n", "\n", "# Create a random search object\n", "random_rf_class = RandomizedSearchCV(\n", " estimator=RandomForestClassifier(n_estimators=80),\n", " param_distributions=param_grid, n_iter=5,\n", " scoring='roc_auc', n_jobs=4, cv=3, refit=True, return_train_score=True\n", ")\n", "\n", "# Fit to the training data\n", "random_rf_class.fit(X_train, y_train)\n", "\n", "# Print the values used for both hyperparameters\n", "print(random_rf_class.cv_results_['param_max_depth'])\n", "print(random_rf_class.cv_results_['param_max_features'])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Comparing Grid and Random Search\n", "- Comparision\n", "\n", "| Grid Search | Random Search |\n", "| ----------- | ------------- |\n", "| Exhaustively tries all combinations within the sample space | Random Selects a subset of combinations within the sample space (that you must specify) |\n", "| No sampling methodology | Can select a sampling methodology (other than uniform) |\n", "| More computationally expensive | Less computationally expensive |\n", "| Guaranteed to find the best score in the sample space | Not guaranteed to find the best score in the sample space (but likely to find a good one fater) |" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Grid and Random Search Side by Side\n", "Visualizing the search space of random and grid search together allows you to easily see the coverage that each technique has and therefore brings to life their specific advantages and disadvantages.\n", "\n", "In this exercise, you will sample hyperparameter combinations in a grid search way as well as a random search way, then plot these to see the difference." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [], "source": [ "def visualize_search(grid_combinations_chosen, random_combinations_chosen):\n", " grid_y, grid_x = [x[0] for x in grid_combinations_chosen], [x[1] for x in grid_combinations_chosen]\n", " rand_y, rand_x = [x[0] for x in random_combinations_chosen], [x[1] for x in random_combinations_chosen]\n", "\n", " # Plot all together\n", " plt.scatter(grid_y + rand_y, grid_x + rand_x, c=['red']*300 + ['blue']*300)\n", " plt.gca().set(xlabel='learn_rate', ylabel='min_samples_leaf', title='Grid and Random Search Hyperparameters')\n", " plt.gca().set_xlim([0.01, 3.0])\n", " plt.gca().set_ylim([5, 25])" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "learn_rate_list = np.linspace(0.01, 3.0, 200)\n", "min_samples_leaf_list = range(5, 25)\n", "\n", "combinations_list = [list(x) for x in product(learn_rate_list, min_samples_leaf_list)]\n", "\n", "# Sample grid coordinates\n", "grid_combinations_chosen = combinations_list[0:300]\n", "\n", "# Create a list of sample indexes\n", "sample_indexes = list(range(0, len(combinations_list)))\n", "\n", "# Randomly sample 300 indexes\n", "random_indexes = np.random.choice(sample_indexes, 300, replace=False)\n", "\n", "# Use indexes to create random sample\n", "random_combinations_chosen = [combinations_list[index] for index in random_indexes]\n", "\n", "# Call the function to produce the visualization\n", "visualize_search(grid_combinations_chosen, random_combinations_chosen)" ] } ], "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": "3.7.6" } }, "nbformat": 4, "nbformat_minor": 4 }