Model-based Clustering using Automatic Differentiation: Confronting Misspecification and High-Dimensional Data
This addresses clustering challenges in statistical modeling for practitioners, but it is incremental as it builds on existing optimization methods with a new penalty.
The paper tackles model-based clustering with Gaussian Mixture Models under misspecification and high-dimensional data, finding that EM outperforms gradient descent in misspecified cases while gradient descent is better for high-dimensional data, and proposes a penalized likelihood approach using a Kullback-Leibler divergence penalty to improve cluster interpretation, showing efficacy in experiments.
We study two practically important cases of model based clustering using Gaussian Mixture Models: (1) when there is misspecification and (2) on high dimensional data, in the light of recent advances in Gradient Descent (GD) based optimization using Automatic Differentiation (AD). Our simulation studies show that EM has better clustering performance, measured by Adjusted Rand Index, compared to GD in cases of misspecification, whereas on high dimensional data GD outperforms EM. We observe that both with EM and GD there are many solutions with high likelihood but poor cluster interpretation. To address this problem we design a new penalty term for the likelihood based on the Kullback Leibler divergence between pairs of fitted components. Closed form expressions for the gradients of this penalized likelihood are difficult to derive but AD can be done effortlessly, illustrating the advantage of AD-based optimization. Extensions of this penalty for high dimensional data and for model selection are discussed. Numerical experiments on synthetic and real datasets demonstrate the efficacy of clustering using the proposed penalized likelihood approach.