How Usable is Automated Feature Engineering for Tabular Data?
Not on the map. Why?
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.
Key facts. Cited by 1 paper 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
24 of 114 references.
References · 24
- 2014OpenML: networked science in machine learningcited by 526
- 2020AutoGluon-Tabular: Robust and Accurate AutoML for Structured Datacited by 383
- 2023Large Language Models for Automated Data Science: Introducing CAAFE for Context-Aware Automated Feature Engineeringcited by 127
- 2019The autofeat Python Library for Automated Feature Engineering and Selectioncited by 73
- 2017Hyperparameter Importance Across Datasetscited by 63
- 2022AMLB: an AutoML Benchmarkcited by 55
- 2022OpenFE: Automated Feature Generation with Expert-level Performancecited by 52
- 2019AutoCross: Automatic Feature Crossing for Tabular Data in Real-World Applicationscited by 40
- 2017Relief-Based Feature Selection: Introduction and Reviewcited by 38
- 2020Feature Interaction Interpretability: A Case for Explaining Ad-Recommendation Systems via Neural Interaction Detectioncited by 32
- 2021LightAutoML: AutoML Solution for a Large Financial Services Ecosystemcited by 32
- 2022AutoField: Automating Feature Selection in Deep Recommender Systemscited by 30
12 more
- 2019Feature Generation by Convolutional Neural Network for Click-Through Rate Predictioncited by 26
- 2020AutoFS: Automated Feature Selection via Diversity-aware Interactive Reinforcement Learningcited by 24
- 2023A Performance-Driven Benchmark for Feature Selection in Tabular Deep Learningcited by 24
- 2020DIFER: Differentiable Automated Feature Engineeringcited by 16
- 2020Simplifying Reinforced Feature Selection via Restructured Choice Strategy of Single Agentcited by 15
- 2024A Data-Centric Perspective on Evaluating Machine Learning Models for Tabular Datacited by 15
- 2023Optimizing Feature Set for Click-Through Rate Predictioncited by 13
- 2024Large Language Models Engineer Too Many Simple Features For Tabular Datacited by 10
- 2016Good Enough Practices in Scientific Computingcited by 9
- 2022LPFS: Learnable Polarizing Feature Selection for Click-Through Rate Predictioncited by 4
- 2021AEFE: Automatic Embedded Feature Engineering for Categorical Featurescited by 2
- 2022MACFE: A Meta-learning and Causality Based Feature Engineering Frameworkcited by 2
Built on by
Citing papers per year: 2026: 1.
1 paper.