The family of power series cure rate models provides a flexible modeling framework for survival data of populations with a cure fraction. In this work, we present a simplified estimation procedure for the maximum like...
详细信息
The family of power series cure rate models provides a flexible modeling framework for survival data of populations with a cure fraction. In this work, we present a simplified estimation procedure for the maximum likelihood (ML) approach. ML estimates are obtained via the expectation-maximization (em) algorithm where the expectation step involves computation of the expected number of concurrent causes for each individual. It has the big advantage that the maximization step can be decomposed into separate maximizations of two lower-dimensional functions of the regression and survival distribution parameters, respectively. Two simulation studies are performed: the first to investigate the accuracy of the estimation procedure for different numbers of covariates and the second to compare our proposal with the direct maximization of the observed log-likelihood function. Finally, we illustrate the technique for parameter estimation on a dataset of survival times for patients with malignant melanoma.
This paper investigates a parameter estimation problem for batch processes through the maximum likelihood method. In batch processes, the initial state usually relates to the states of previous batches. The proposed a...
详细信息
This paper investigates a parameter estimation problem for batch processes through the maximum likelihood method. In batch processes, the initial state usually relates to the states of previous batches. The proposed algorithm takes batch-to-batch correlations into account by employing an initial state transition equation to model the dynamics along the batch dimension. By treating the unmeasured states and the parameters as hidden variables, the maximum likelihood estimation is accomplished through the expectation-maximization (em) algorithm, where the smoothing for the terminal state and the filtering for the initial state are specially considered. Due to the nonlinearity and non-Gaussianity in the state space model, particle filtering methods are employed for the implementation of filtering and smoothing. Through alternating between the expectation step and the maximization step, the unknown parameters along with states are estimated. Simulation examples demonstrate the proposed estimation approach. (C) 2013 Elsevier Ltd. All rights reserved.
In this work, a probabilistic model is established for recurrent networks, The em (expectation-maximization) algorithm is then applied to derive a new fast training algorithm for recurrent networks through mean-field ...
详细信息
In this work, a probabilistic model is established for recurrent networks, The em (expectation-maximization) algorithm is then applied to derive a new fast training algorithm for recurrent networks through mean-field approximation, This new algorithm converts training a complicated recurrent network into training an array of individual feedforward neurons, These neurons are then trained via a linear weighted regression algorithm. The training time has been improved by five to 15 times on benchmark problems.
This paper describes how 3D facial pose may be estimated by fitting a template to 2D feature locations. The fitting process is realised as projecting the control points of a 3D template onto the 2D feature locations u...
详细信息
This paper describes how 3D facial pose may be estimated by fitting a template to 2D feature locations. The fitting process is realised as projecting the control points of a 3D template onto the 2D feature locations under orthographic projection. The parameters of the orthographic projection are iteratively estimated using the em algorithm. The method is evaluated on both contrived data with known ground-truth together with some more naturalistic imagery. These experiment, reveal that under favourable conditions the algorithm can estimate facial pitch to within 3degrees. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.
The em algorithm, e.g., the Baum-Welch re-estimation, is an important tool for parameter estimation in discrete-time hidden Markov models. We present a direct re-estimation of rate constants for applications in which ...
详细信息
The em algorithm, e.g., the Baum-Welch re-estimation, is an important tool for parameter estimation in discrete-time hidden Markov models. We present a direct re-estimation of rate constants for applications in which the underlying Marker process is continuous in time. Previous estimation of discrete-time transition probabilities is not necessary.
We propose a genetic-based expectation-maximization (GA-em) algorithm for learning Gaussian mixture models from multivariate data. This algorithm is capable of selecting the number of components of the model using the...
详细信息
We propose a genetic-based expectation-maximization (GA-em) algorithm for learning Gaussian mixture models from multivariate data. This algorithm is capable of selecting the number of components of the model using the minimum description length (MDL) criterion. Our approach benefits from the properties of Genetic algorithms (GA) and the em algorithm by combination of both into a single procedure. The population-based stochastic search of the GA explores the search space more thoroughly than the em method. Therefore, our algorithm enables escaping from local optimal solutions since the algorithm becomes less sensitive to its initialization. The GA-em algorithm is elitist which maintains the monotonic convergence property of the em algorithm. The experiments on simulated and real data show that the GA-em outperforms the em method since: 1) We have obtained a better MDL score while using exactly the same termination condition for both algorithms. 2) Our approach identifies the number of components which were used to generate the underlying data more often than the em algorithm.
Maximum likelihood is an important approach to analysis of two-level structural equation models. Different algorithms for this purpose have been available in the literature. In this paper, we present a new formulation...
详细信息
Maximum likelihood is an important approach to analysis of two-level structural equation models. Different algorithms for this purpose have been available in the literature. In this paper, we present a new formulation of two-level structural equation models and develop an em algorithm for fitting this formulation. This new formulation covers a variety of two-level structural equation models. As a result,the proposed em algorithm is widely applicable in practice. A practical example illustrates the performance of the em algorithm and the maximum likelihood statistic.
In this paper we describe a new method for the analysis of electron microscope autoradiographs. It uses a Poisson model to describe the autoradiographic grain distribution, and the method of maximum likelihood to esti...
详细信息
In this paper we describe a new method for the analysis of electron microscope autoradiographs. It uses a Poisson model to describe the autoradiographic grain distribution, and the method of maximum likelihood to estimate the radioactive intensities. An iterative procedure is derived from the em algorithm to produce maximum likelihood estimates. The procedure leads to simple iterative calculations and a complete treatment of edge effects. The resulting method of analysis enables data from the whole area of autoradiograph plates to be utilized without the need for any form of windowing to guard against edge effects. It also facilitates inference about the precision of estimated intensities and about comparisons between intensities. The estimation procedures are illustrated with real autoradiograph data from a study of rabbit blood plasma.
Localization technologies play an increasingly important role in pervasive applications of wireless sensor networks. Since the number of targets is usually limited, localization benefits from compressed sensing (CS): ...
详细信息
Localization technologies play an increasingly important role in pervasive applications of wireless sensor networks. Since the number of targets is usually limited, localization benefits from compressed sensing (CS): measurements number can be greatly reduced. Despite many CS-based localization schemes, existing solutions implicitly assume that all targets fall on a fixed grid exactly. When the assumption is violated, the mismatch between the assumed and actual sparsifying dictionaries can deteriorate the localization performance significantly. To address such a problem, in this paper, we propose a novel and iterative multiple target counting and localization framework. The key idea behind the framework is to dynamically adjust the grid to alleviate or even eliminate dictionary mismatch. The contribution of this paper is twofold. First, we consider the off-grid target issue in CS-based localization and formulate multiple target counting and localization as a joint sparse signal recovery and parameter estimation problem. Second, we solve the joint optimization problem using a variational Bayesian expectation-maximization algorithm where the sparse signal and parameter are iteratively updated in the variational Bayesian expectation-step and variational Bayesian maximization-step, respectively. Extensive simulation results highlight the superior performance of the proposed framework in terms of probability of correct counting and average localization error.
An expectation-maximization (em) algorithm for learning sparse and overcomplete representations is presented in this paper. We show that the estimation of the conditional moments of the posterior distribution can be a...
详细信息
An expectation-maximization (em) algorithm for learning sparse and overcomplete representations is presented in this paper. We show that the estimation of the conditional moments of the posterior distribution can be accomplished by maximum a posteriori estimation. The approximate conditional moments,enable the development of an em algorithm for learning the overcomplete basis vectors and inferring the most probable basis coefficients. (C) 2003 Elsevier B.V. All rights reserved.
暂无评论