A Dirichlet-Multinomial mixture model of Statistical Science: Mapping the shift of a paradigm

Massimo Bilancia and Rade Dačević

Journal of Informetrics, 2025, vol. 19, issue 1

Abstract: Using Bayesian natural language processing (NLP) methods and a scalable variational algorithm tailored for mixtures of discrete positive data, we analyzed a large corpus of 111,411 eprints submitted to the arXiv repository between 1994 and 2022 in the Statistics category (the primary classification for these eprints on arXiv). Our objective is to assess the impact of Machine Learning (ML) on the field of Statistics–specifically, to determine whether the introduction of ML has led to a fundamental paradigm shift, transforming traditional statistical problems or creating entirely new ones, or if this perceived revolution is primarily occurring outside the field of Statistics. Our findings suggest that the only significant paradigm shift for Statistics as a scientific discipline remains the Bayesian revolution that began in the early 1990s.

Keywords: Statistical Science; Natural language processing; Dirichlet-Multinomial mixture models; Bayesian statistics; Variational inference (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S1751157724001457
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:infome:v:19:y:2025:i:1:s1751157724001457

DOI: 10.1016/j.joi.2024.101633

Access Statistics for this article

Journal of Informetrics is currently edited by Leo Egghe

More articles in Journal of Informetrics from Elsevier
Bibliographic data for series maintained by Catherine Liu ().