Before we can validate models, we need an understanding of how to create and work with them. This chapter provides an introduction to running regression and classification models in scikit-learn. We will use this model building foundation throughout the remaining chapters. 

Introduction to model validation

Modeling steps

Seen vs. unseen data

Regression models

Set parameters and fit a model

Feature importances

Classification models

Classification predictions

Reusing model parameters

Random forest classifier

Basic Modeling in scikit-learn

This chapter focuses on the basics of model validation. From splitting data into training, validation, and testing datasets, to creating an understanding of the bias-variance tradeoff, we build the foundation for the techniques of K-Fold and Leave-One-Out validation practiced in chapter three. 

Creating train, test, and validation datasets

Create one holdout set

Create two holdout sets

Why use holdout sets

Accuracy metrics: regression models

Mean absolute error

Mean squared error

Performance on data subsets

Classification metrics

Confusion matrices

Confusion matrices, again

Precision vs. recall

The bias-variance tradeoff

Error due to under/over-fitting

Am I underfitting?

Validation Basics

Holdout sets are a great start to model validation. However, using a single train and test set if often not enough. Cross-validation is considered the gold standard when it comes to validating model performance and is almost always used when tuning model hyper-parameters. This chapter focuses on performing cross-validation to validate model performance. 

The problems with holdout sets

Two samples

Potential problems

Cross-validation

scikit-learn's KFold()

Using KFold indices

sklearn's cross_val_score()

scikit-learn's methods

Implement cross_val_score()

Leave-one-out-cross-validation (LOOCV)

When to use LOOCV

Leave-one-out-cross-validation

Cross Validation

The first three chapters focused on model validation techniques. In chapter 4 we apply these techniques, specifically cross-validation, while learning about hyperparameter tuning. After all, model validation makes tuning possible and helps us select the overall best model. 

Introduction to hyperparameter tuning

Creating Hyperparameters

Running a model using ranges

RandomizedSearchCV

Preparing for RandomizedSearch

Implementing RandomizedSearchCV

Selecting your final model

Best classification accuracy

Selecting the best precision model

Course completed!

Selecting the best model with Hyperparameter tuning. 

Candy dataset

Tic-Tac-Toe dataset

Machine learning models are easier to implement now more than ever before. Without proper validation, the results of running new data through a model might not be as accurate as expected. Model validation allows analysts to confidently answer the question, how good is your model? We will answer this question for classification models using the complete set of tic-tac-toe endgame scenarios, and for regression models using fivethirtyeight’s ultimate Halloween candy power ranking dataset. In this course, we will cover the basics of model validation, discuss various validation techniques, and begin to develop tools for creating validated and high performing models.

Supervised Learning with scikit-learn

Learn the basics of model validation,  validation techniques, and begin creating validated and high performing models.

Model Validation in Python

Scientifique en apprentissage automatique en Python

Why use holdout sets

Model Validation in Python

Exercice interactif pratique