Evaluating Large Language Models for Automated Data Cleaning and Feature Engineering in Clinical Datasets
Shimul A Babli, Pulkit Gairola, Chukwuemeka E Ogbu, Francesco Alessi Longa, Aditya Rautaray, Opeyemi S Alamu, Syed Faheem Haider Rizvi · 2026
Read Evaluating Large Language Models for Automated Data Cleaning and Feature Engineering in Clinical Datasets by Shimul A Babli, Pulkit Gairola, Chukwuemeka E Ogbu, Francesco Alessi Longa, Aditya Rautaray, Opeyemi S Alamu, Syed Faheem Haider Rizvi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13245537.txt.
Description
Shimul Babli and co-authors tested large language models at cleaning clinical data in two thousand twenty-six. Health records hold missing values, outliers, and mixed coding, and hand-built cleaning rules take much expert time. The authors asked if chat-style models could write the cleaning code instead.
They tested three models on five cleaning tasks and two feature-building tasks across three public datasets. One model scored highest on cleaning, and all three beat the rule-based baseline. Data cleaned by the top model also gave the best death prediction in one dataset. New features from the models added small gains. Expert checks remain essential, the authors warn.
Catalog record
| CREATOR | Shimul A Babli, Pulkit Gairola, Chukwuemeka E Ogbu, Francesco Alessi Longa, Aditya Rautaray, Opeyemi S Alamu, Syed Faheem Haider Rizvi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |