The tidymodels framework in R is a collection of packages for modeling.

Within tidymodels, the parsnip package is primarily responsible for specifying models in a way that is independent of the underlying modeling engines. The set_engine() function in parsnip allows users to specify which computational engine to use for modeling, enabling the same model specification to be used across different packages and implementations.

In this blog series, we will look at some commonly used models and engines within the tidymodels package

1. Linear Regression (lm): The classic linear regression model, with the default engine being stats, referring to the base R stats package.
2. Logistic Regression (logistic_reg): Used for binary classification problems, with engines like stats for the base R implementation and glmnet for regularized regression.
3. Random Forest (rand_forest): A popular ensemble method for classification and regression tasks, with engines like ranger and randomForest.
4. Boosted Trees (boost_tree): Used for boosting tasks, with engines such as xgboost, lightgbm, and catboost.
5. Decision Trees (decision_tree): A base model for classification and regression, with engines like rpart and C5.0.
6. K-Nearest Neighbors (nearest_neighbor): A simple yet effective non-parametric method, with engines like kknn and caret.
7. Principal Component Analysis (pca): For dimensionality reduction, with the stats engine.
8. Lasso and Ridge Regression (linear_reg): For regression with regularization, specifying the penalty parameter and using engines like glmnet.

Click here for some resources I found:

1. https://rviews.rstudio.com/2019/06/19/a-gentle-intro-to-tidymodels
2. https://rpubs.com/chenx/tidymodels_tutorial
3. https://bookdown.org/paul/ai_ml_for_social_scientists/06_01_ml_with_tidymodels.html