Asymptotic properties of mixture-of-experts models
نویسندگان
چکیده
Abstract. The statistical properties of the likelihood ratio test statistic (LRTS) for mixture-of-expert models are addressed in this paper. This question is essential when estimating the number of experts in the model. Our purpose is to extend the existing results for mixtures (Liu and Shao, 2003) and mixtures of multilayer perceptrons (Olteanu and Rynkiewicz, 2008). In this paper we study a simple example which embodies all the difficulties arising in such models. We find that in some cases the LRTS diverges but, with additional assumptions, the behavior of such models can be totally explicated.
منابع مشابه
Asymptotic Analysis of Binary Gas Mixture Separation by Nanometric Tubular Ceramic Membranes: Cocurrent and Countercurrent Flow Patterns
Analytical gas-permeation models for predicting the separation process across membranes (exit compositions and area requirement) constitutes an important and necessary step in understanding the overall performance of membrane modules. But, the exact (numerical) solution methods suffer from the complexity of the solution. Therefore, solutions of nonlinear ordinary differential equations th...
متن کاملEstimation in ARMA models based on signed ranks
In this paper we develop an asymptotic theory for estimation based on signed ranks in the ARMA model when the innovation density is symmetrical. We provide two classes of estimators and we establish their asymptotic normality with the help of the asymptotic properties for serial signed rank statistics. Finally, we compare our procedure to the one of least-squares, and we illustrate the performa...
متن کاملDetermination of the Rheological Properties of Hydroxyl Terminated Polybutadiene (HTPB) Mixtures With Energetic Materials and Presenting Empricial Models
Rheological Properties Such as Viscosity (η), Shear Stress (τ), and Torque (M) of the mixtures of (HTPB) with Octagon (HMX), Hexogen (RDX), and 2, 6 Diamino-4-Phenyl-1, 3, 5 Triazine (DAPTA) mixtures were measured. The experimental design was arranged for three factors at two levels (High and low levels). Temperature of the mixture (°C), Speed of the stirrer (rpm), Mixing Time (minutes) have be...
متن کاملSkew Laplace Finite Mixture Modelling
‎This paper presents a new mixture model via considering the univariate skew Laplace distribution‎. ‎The new model can handle both heavy tails and skewness and is multimodal‎. ‎Describing some properties of the proposed model‎, ‎we present a feasible EM algorithm for iteratively‎ ‎computing maximum likelihood estimates‎. ‎We also derive the observ...
متن کاملAsymptotic Optimality of the Bayes Estimator on Differentiable in Quadratic Mean Models
This paper deals with the study of the Bayes estimator’s asymptotic properties on Differentiable in Quadratic Mean (DQM) models in the case of independent and identically distributed observations. The investigation is led in order to define weak assumptions on the model under which this estimator is asymptotically efficient, regular and asymptotically of minimal risk. The results of the paper a...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Neurocomputing
دوره 74 شماره
صفحات -
تاریخ انتشار 2010