How Usable is Automated Feature Engineering for Tabular Data?

2025-08-19

AI Summary

This work addresses the usability bottleneck of automated feature engineering (AutoFE) for tabular data. We conduct the first empirical evaluation of 53 mainstream AutoFE methods, assessing practical dimensions including ease of deployment, documentation completeness, community activity, and support for user-defined resource constraints (e.g., time and memory budgets). Results reveal that most tools suffer from complex installation, incomplete or outdated documentation, inactive maintenance, and—critically—lack configurable resource limits, severely hindering industrial adoption. To bridge this gap, we propose a “high-usability”-oriented AutoFE evaluation framework that prioritizes resource controllability and end-user friendliness in engineering design. Our findings expose a substantial practicality gap in current AutoFE tooling, providing empirical evidence and concrete guidance for developing next-generation AutoFE systems that are trustworthy, configurable, and sustainably maintainable.

Technology Category

• Machine Learning: Auto ML and Hyperparameter Tuning
• Humans and AI: User Experience and Usability
• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

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.

Problem

• Investigating usability of automated feature engineering methods
• Evaluating practicality for practitioners in machine learning
• Assessing documentation and community support for AutoFE

Innovation

• Automated feature engineering methods evaluation
• Usability assessment with time constraints
• Survey highlighting need for improved automation