Learn

Learn how to go farther with tidymodels in your modeling and machine learning projects.

After you know what you need to get started with tidymodels, you can learn more and go further. Find articles here to help you solve specific problems using the tidymodels framework.

A predictive modeling case study

Develop, from beginning to end, a predictive model using best practices.

Accounting for Censoring in Performance Metrics for Event Time Data

statistical analysis

survival analysis

Learn how tidymodels uses causal inference tools to measure performance of survival models.

An introduction to calibration with tidymodels

Learn how the probably package can improve classification and regression models.

Are GPT detectors fair? A machine learning fairness case study

Releases of language models are often soon followed up with detectors, claiming to detect whether text was written by a human or a given language model. A 2023 study argues…

Bootstrap resampling and tidy regression models

statistical analysis

tidying results

confidence intervals

Apply bootstrap resampling to estimate uncertainty in model parameters.

Build a model

tidying results

Get started by learning how to specify and train a model using tidymodels.

Classification models using a neural network

neural networks

Train a classification model and evaluate its performance.

Confidence Intervals for Performance Metrics

confidence intervals

tidymodels has high-level functions to produce bootstrap confidence intervals for performance statistics.

Conformal inference for regression models

The probably package has functions to create prediction intervals for regression models.

Correlation and regression fundamentals with tidy data principles

statistical analysis

tidying results

Analyze the results of correlation tests and simple regression models for many data sets at once.

Create your own broom tidier methods

developer tools

Write tidy(), glance(), and augment() methods for new model objects.

Create your own recipe step function

developer tools

Write a new recipe step for data preprocessing.

Creating case weights based on time

Create models that use coefficients, extract them from fitted models, and visualize them.

Custom performance metrics

developer tools

Create a new performance metric and integrate it with yardstick functions.

Dynamic Performance Metrics for Event Time Data

statistical analysis

survival analysis

Let’s discuss how to compute modern performance metrics for time-to-event models.

Evaluate your model with resampling

Measure model performance by generating different versions of the training data through resampling.

Fair prediction of hospital readmission: a machine learning fairness case study

With information on a diabetes patient’s stay at a hospital like demographics, diagnostic results, payment, and medications, a hospital can train a machine learning model to…

How long until building complaints are dispositioned? A survival analysis case study

statistical analysis

survival analysis

Learn how to use tidymodels for survival analysis.

How to build a parsnip model

developer tools

Create a parsnip model function from an existing model implementation.

How to create a tuning parameter function

developer tools

Build functions to use in tuning both quantitative and qualitative parameters.

Hypothesis testing using resampling and tidy data

statistical analysis

hypothesis testing

Perform common hypothesis tests for statistical inference using flexible functions.

Iterative Bayesian optimization of a classification model

Identify the best hyperparameters for a model using Bayesian optimization of iterative search.

K-means clustering with tidy data principles

statistical analysis

tidying results

Summarize clustering characteristics and estimate the best number of clusters for a data set.

Model tuning via grid search

Choose hyperparameters for a model by training on a grid of many possible parameter values.

Modeling time series with tidy resampling

Calculate performance estimates for time series forecasts using resampling.

Multivariate analysis using partial least squares

multivariate analysis

Build and fit a predictive model with more than one outcome.

Nested resampling

Estimate the best hyperparameters for a model using nested resampling.

Preprocess your data with recipes

tidying results

Prepare data for modeling with modular preprocessing steps.

Regression models two ways

Create and train different kinds of regression models with different computational engines.

Statistical analysis of contingency tables

statistical analysis

hypothesis testing

Use tests of independence and goodness of fit to analyze tables of counts.

Subsampling for class imbalances

Improve model performance in imbalanced data sets through undersampling or oversampling.

Tune model parameters

Estimate the best values for hyperparameters that cannot be learned directly during model training.

Tuning text models

Prepare text data for predictive modeling and tune with both grid and iterative search.

Working with model coefficients

tidying results

Create models that use coefficients, extract them from fitted models, and visualize them.