Evolutionary clustering for categorical data using parametric links among multinomial mixture models

Abstract

A novel evolutionary clustering method for temporal categorical data based on parametric links among the Multinomial mixture models is proposed. Besides clustering, the main goal is to interpret the evolution of clusters over time. To this aim, first the formulation of a generalized model that establishes parametric links among two Multinomial mixture models is proposed. Afterward, different parametric sub-models are defined in order to model the typical evolution of the clustering structure. Model selection criteria allow to select the best sub-model and thus to guess the clustering evolution. For the experiments, the proposed method is first evaluated with synthetic temporal data. Next, it is applied to analyze the annotated social media data. Results show that the proposed method is better than the state-of-the-art based on the common evaluation metrics. Additionally, it can provide interpretation about the temporal evolution of the clusters.

A hierarchical mixture cure model with unobserved heterogeneity for credit risk

“…Finally, alternative distributions can be used to model the subgroup densities, e.g., the skew-t distribution employed in Murray et al (2017), and the approach of Hasnat et al (2017) can be used to investigate the evolution of subgroups over time. {π * j (z) exp −φ * j (x)H * j 0 (t) − π * j (z)}, it follows that π j (z) = π * j (z), φ j (x) = φ * j (x), H j 0 = H * j 0 , for j = 1, 2.…”