On generalized Gower distance for mixed-type data: extensive simulation study and new software tools

Data scientists address real-world problems using multivariate and heterogeneous datasets, characterized by multiple variables of different natures. Selecting a suitable distance function between units is crucial, as many statistical techniques and machine learning algorithms depend on this concept. Traditional distances, such as Euclidean or Manhattan, are unsuitable for mixed-type data, and although the Gower distance was designed to handle this kind of data, it may lead to suboptimal results in the presence of outlying units or an underlying correlation structure. This paper defines and explores robust distances for mixed-type data, namely robust generalized Gower and robust related metric scaling. The Python package called robust_mixed_dist is developed, which enables the computation of these robust proposals as well as classical ones.

Recommended citation: Grané, A., & Scielzo-Ortiz, F. (2025). "Robust distances for mixed-type data: Robust generalized Gower and robust related metric scaling." SORT - Statistics and Operations Research Transactions. https://doi.org/10.57645/20.8080.02.28
Download Paper