# How Usable is Automated Feature Engineering for Tabular Data?

arXiv: https://arxiv.org/abs/2508.13932 · HTML: https://mlsanity.com/paper/2508.13932/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 114 references that are in the map.
- First posted on arXiv 19 Aug 2025.

## Builds on (most cited first, 24 of 24)

- [OpenML: networked science in machine learning](https://mlsanity.com/paper/1407.7722/) (2014)
- [AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data](https://mlsanity.com/paper/2003.06505/) (2020)
- [Large Language Models for Automated Data Science: Introducing CAAFE for Context-Aware Automated Feature Engineering](https://mlsanity.com/paper/2305.03403/) (2023)
- [The autofeat Python Library for Automated Feature Engineering and Selection](https://mlsanity.com/paper/1901.07329/) (2019)
- [Hyperparameter Importance Across Datasets](https://mlsanity.com/paper/1710.04725/) (2017)
- [AMLB: an AutoML Benchmark](https://mlsanity.com/paper/2207.12560/) (2022)
- [OpenFE: Automated Feature Generation with Expert-level Performance](https://mlsanity.com/paper/2211.12507/) (2022)
- [AutoCross: Automatic Feature Crossing for Tabular Data in Real-World Applications](https://mlsanity.com/paper/1904.12857/) (2019)
- [Relief-Based Feature Selection: Introduction and Review](https://mlsanity.com/paper/1711.08421/) (2017)
- [Feature Interaction Interpretability: A Case for Explaining Ad-Recommendation Systems via Neural Interaction Detection](https://mlsanity.com/paper/2006.10966/) (2020)
- [LightAutoML: AutoML Solution for a Large Financial Services Ecosystem](https://mlsanity.com/paper/2109.01528/) (2021)
- [AutoField: Automating Feature Selection in Deep Recommender Systems](https://mlsanity.com/paper/2204.09078/) (2022)
- [Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction](https://mlsanity.com/paper/1904.04447/) (2019)
- [AutoFS: Automated Feature Selection via Diversity-aware Interactive Reinforcement Learning](https://mlsanity.com/paper/2008.12001/) (2020)
- [A Performance-Driven Benchmark for Feature Selection in Tabular Deep Learning](https://mlsanity.com/paper/2311.05877/) (2023)
- [DIFER: Differentiable Automated Feature Engineering](https://mlsanity.com/paper/2010.08784/) (2020)
- [Simplifying Reinforced Feature Selection via Restructured Choice Strategy of Single Agent](https://mlsanity.com/paper/2009.09230/) (2020)
- [A Data-Centric Perspective on Evaluating Machine Learning Models for Tabular Data](https://mlsanity.com/paper/2407.02112/) (2024)
- [Optimizing Feature Set for Click-Through Rate Prediction](https://mlsanity.com/paper/2301.10909/) (2023)
- [Large Language Models Engineer Too Many Simple Features For Tabular Data](https://mlsanity.com/paper/2410.17787/) (2024)
- [Good Enough Practices in Scientific Computing](https://mlsanity.com/paper/1609.00037/) (2016)
- [LPFS: Learnable Polarizing Feature Selection for Click-Through Rate Prediction](https://mlsanity.com/paper/2206.00267/) (2022)
- [AEFE: Automatic Embedded Feature Engineering for Categorical Features](https://mlsanity.com/paper/2110.09770/) (2021)
- [MACFE: A Meta-learning and Causality Based Feature Engineering Framework](https://mlsanity.com/paper/2207.04010/) (2022)

## Built on by (most cited first, 1 of 1)

- [TabPrep: Closing the Feature Engineering Gap in Tabular Benchmarks](https://mlsanity.com/paper/2606.02384/) (2026)

## Abstract (arXiv)

Tabular data, consisting of rows and columns, is omnipresent across various machine learning applications. Each column represents a feature, and features can be combined or transformed to create new, more informative features. Such feature engineering is essential to achieve peak performance in machine learning. Since manual feature engineering is expensive and time-consuming, a substantial effort has been put into automating it. Yet, existing automated feature engineering (AutoFE) methods have never been investigated regarding their usability for practitioners. Thus, we investigated 53 AutoFE methods. We found that these methods are, in general, hard to use, lack documentation, and have no active communities. Furthermore, no method allows users to set time and memory constraints, which we see as a necessity for usable automation. Our survey highlights the need for future work on usable, well-engineered AutoFE methods.