Streamlines machine learning data preprocessing in R using standardized Tidymodels recipes patterns.
This skill provides a comprehensive library of patterns for the R 'recipes' package, enabling data scientists to build robust, reproducible preprocessing pipelines. It covers the entire feature engineering lifecycle, including numeric normalization, categorical encoding, missing data imputation, and dimensionality reduction. By following the included best-practice guidelines for step ordering, users can prevent information leakage and ensure their models generalize well. It also integrates specialized extensions for handling text data, date/time features, and class imbalance, making it an essential tool for Tidymodels practitioners.
主要功能
01Best-practice step ordering to prevent data leakage during preprocessing
02Standardized patterns for numeric normalization and non-linear transformations
03Specialized steps for text processing, date/time extraction, and class imbalance
04Advanced missing data imputation using KNN, Bagging, and Linear models
050 GitHub stars
06Comprehensive categorical encoding including dummy variables and target encoding
使用场景
01Handling complex datasets with high-cardinality categories or missing values
02Implementing automated feature selection and dimensionality reduction like PCA or UMAP
03Building production-grade preprocessing pipelines for Tidymodels workflows
