{"id":"arxiv/2602.16540","url":"https://arxiv.org/abs/2602.16540","tags":["stat.ME","math.ST"],"score":0,"title":"Generalised Linear Models Driven by Latent Processes: Asymptotic Theory and Applications","domain":"stat.ME","source":"arxiv","authors":["Wagner Barreto-Souza","Ngai Hang Chan"],"reviews":0,"subject":"stat.ME","summary":"This paper introduces a class of generalised linear models (GLMs) driven by latent processes for modelling count, real-valued, binary, and positive continuous time series. Extending earlier latent-process regression frameworks based on Poisson or one-parameter exponential family assumptions, we allow the conditional distribution of the response to belong to a bi-parameter exponential family, with the latent process entering the conditional mean multiplicatively. This formulation substantially broadens the scope of latent-process GLMs, for instance, it naturally accommodates gamma responses…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-02-18","reactions":{}}