We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Fairness implications of encoding protected categorical attributes

Abstract: Protected attributes are often presented as categorical features that need to be encoded before feeding them into a machine learning algorithm. Encoding these attributes is paramount as they determine the way the algorithm will learn from the data. Categorical feature encoding has a direct impact on the model performance and fairness. In this work, we compare the accuracy and fairness implications of the two most well-known encoders: one-hot encoding and target encoding. We distinguish between two types of induced bias that can arise while using these encodings and can lead to unfair models. The first type, irreducible bias, is due to direct group category discrimination and a second type, reducible bias, is due to large variance in less statistically represented groups. We take a deeper look into how regularization methods for target encoding can improve the induced bias while encoding categorical features. Furthermore, we tackle the problem of intersectional fairness that arises when mixing two protected categorical features leading to higher cardinality. This practice is a powerful feature engineering technique used for boosting model performance. We study its implications on fairness as it can increase both types of induced bias
Comments: 22 pages
Subjects: Machine Learning (cs.LG); Computers and Society (cs.CY); Data Structures and Algorithms (cs.DS); Machine Learning (stat.ML)
Cite as: arXiv:2201.11358 [cs.LG]
  (or arXiv:2201.11358v1 [cs.LG] for this version)

Submission history

From: Carlos Mougan [view email]
[v1] Thu, 27 Jan 2022 07:39:26 GMT (387kb)

Link back to: arXiv, form interface, contact.