Introducing evoFE: Evolutionary Feature Engineering in R for XGBoost and LightGBM Hey everyone, I’m excited to share a new package I've been working on: evoFE (Evolutionary Feature Engineering). Manually engineering features (creating interaction terms, ratios, group aggregations, clustering, or binning) is one of the most time-consuming parts of building tabular machine learning models. evoFE aims to automate this process by using a Genetic Algorithm (GA) to search the space of possible feature recipes, automatically combining and optimizing transformations to maximize your model's validation score. GitHub Repository: https://github.com/tanopereira/evoFE Documentation Website: https://tanopereira.github.io/evoFE/ Key Features: Hierarchical Feature Chaining: Unlike simpler search tools that only test single-level operations, evoFE can evolve multi-level trees of features. It can learn that log(divide(x1, x2)) or groupby_zscore(umap_1, group_col) is highly predictive and build on top of them over generations.…