Non-standard boundary behaviour in binary mixture models

Consider a binary mixture model of the form , where is standard Gaussian and is a completely specified heavy-tailed distribution with the same support. For a sample of independent and identically distributed values , the maximum likelihood estimator is asymptotically normal provided that is an interior point. This paper investigates the large-sample behaviour for boundary points, which is entirely different and strikingly asymmetric for and . The reason for the asymmetry has to do with typical choices such that is an extreme boundary point and is usually not extreme. On the right boundary, well known results on boundary parameter problems are recovered, giving . On the left boundary, , where indexes the domain of attraction of the density ratio when . For , which is the most important case in practice, we show how the tail behaviour of governs the rate at which tends to zero. A new limit theorem for the joint distribution of the sample maximum and sample mean conditional on positivity establishes multiple inferential anomalies. Most notably, given , the likelihood ratio statistic has a conditional null limit distribution determined by the joint limit theorem. We show through this route that no advantage is gained by extending the single distribution to the nonparametric composite mixture generated by the same tail-equivalence class.
View on arXiv