On Clustering Categories of Categorical Predictors in Generalized Linear Models
This addresses the issue of wasteful encoding and overfitting for practitioners dealing with high-cardinality categorical data, though it is incremental as it builds on existing clustering and modeling techniques.
The paper tackles the problem of high complexity in Generalized Linear Models with categorical predictors by clustering categories to reduce the number of coefficients, resulting in substantial complexity reduction without harming accuracy in real-world datasets.
We propose a method to reduce the complexity of Generalized Linear Models in the presence of categorical predictors. The traditional one-hot encoding, where each category is represented by a dummy variable, can be wasteful, difficult to interpret, and prone to overfitting, especially when dealing with high-cardinality categorical predictors. This paper addresses these challenges by finding a reduced representation of the categorical predictors by clustering their categories. This is done through a numerical method which aims to preserve (or even, improve) accuracy, while reducing the number of coefficients to be estimated for the categorical predictors. Thanks to its design, we are able to derive a proximity measure between categories of a categorical predictor that can be easily visualized. We illustrate the performance of our approach in real-world classification and count-data datasets where we see that clustering the categorical predictors reduces complexity substantially without harming accuracy.