Nonparametric Maximum Likelihood Estimator Research Articles

Nonparametric Benchmark Dose Estimation with Continuous Dose‐Response Data

AbstractWe propose a new method for risk‐analytic benchmark dose (BMD) estimation in a dose‐response setting when the responses are measured on a continuous scale. For each dose leveld, the observationX(d) is assumed to follow a normal distribution:. No specific parametric form is imposed upon the meanμ(d), however. Instead, nonparametric maximum likelihood estimates ofμ(d) andσare obtained under a monotonicity constraint onμ(d). For purposes of quantitative risk assessment, a ‘hybrid’ form of risk function is defined for any dosedasR(d) =P[X(d) <c], wherec> 0 is a constant independent ofd. The BMD is then determined by inverting theadditional risk functionRA(d) =R(d) −R(0) at some specified value of benchmark response. Asymptotic theory for the point estimators is derived, and a finite‐sample study is conducted, using both real and simulated data. When a large number of doses are available, we propose an adaptive grouping method for estimating the BMD, which is shown to have optimal mean integrated squared error under appropriate designs.

An Exponential Tilt Mixture Model for Time-to-Event Data to Evaluate Treatment Effect Heterogeneity in Randomized Clinical Trials.

Evaluating the effect of a treatment on a time-to-event outcome is the focus of many randomized clinical trials. It is often observed that the treatment effect is heterogeneous, where only a subgroup of the patients may respond to the treatment due to some unknown mechanism such as genetic polymorphism. In this paper, we propose a semiparametric exponential tilt mixture model to estimate the proportion of patients who respond to the treatment and to assess the treatment effect. Our model is a natural extension of parametric mixture models to a semiparametric setting with a time-to-event outcome. We propose a nonparametric maximum likelihood estimation approach for inference and establish related asymptotic properties. Our method is illustrated by a randomized clinical trial on biodegradable polymer-delivered chemotherapy for malignant gliomas patients.

A self-consistent estimator of survival function with interval-censored and left-truncated data

Interval censoring refers to a situation in which, Ti∗, the time to occurrence of an event of interest is only known to lie in an interval [Li∗,Ri∗]. In some cases, the variable Ti∗ also suffers left-truncation. The nonparametric maximum likelihood estimator (NPMLE) of the survival function of Ti∗ can be obtained by using an EM algorithm of Turnbull (1976). One disadvantage of the NPMLE is that it is not uniquely defined in the innermost intervals. In this article, we propose a self-consistent estimator (SCE), which does not require interpolation. Furthermore, we show that the NPMLE is also an SCE. We establish the consistency of the SCE under certain conditions, which implies that the NPMLE is also a consistent estimator. A simulation study is conducted to compare the performance between the SCE and the NPMLE.

Quantile residual lifetime with right-censored and length-biased data

Right-censored length-biased data are commonly encountered in many applications such as cancer screening trials, prevalent cohort studies and labor economics. Such data have a unique structure that is different from traditional survival data. In this paper, we propose an estimator of the quantile residual lifetime (QRL) with this kind of data based on the nonparametric maximum likelihood estimation method. In addition, we develop two tests by taking difference and ratio of the QRL from two independent samples. We also establish the asymptotic properties of the proposed estimator and the test statistics. Simulation studies are performed to demonstrate that the proposed estimator works well in finite-sample situations. We illustrate its application using two data examples: one is the Oscars Award data, the other is the Channing house data.

A consistent NPMLE of the joint distribution function with competing risks data under the dependent masking and right-censoring model.

Dinse (Biometrics, 38:417-431, 1982) provides a special type of right-censored and masked competing risks data and proposes a non-parametric maximum likelihood estimator (NPMLE) and a pseudo MLE of the joint distribution function [Formula: see text] with such data. However, their asymptotic properties have not been studied so far. Under the extention of either the conditional masking probability (CMP) model or the random partition masking (RPM) model (Yu and Li, J Nonparametr Stat 24:753-764, 2012), we show that (1) Dinse's estimators are consistent if [Formula: see text] takes on finitely many values and each point in the support set of [Formula: see text] can be observed; (2) if the failure time is continuous, the NPMLE is not uniquely determined, and the standard approach (which puts weights only on one element in each observed set) leads to an inconsistent NPMLE; (3) in general, Dinse's estimators are not consistent even under the discrete assumption; (4) we construct a consistent NPMLE. The consistency is given under a new model called dependent masking and right-censoring model. The CMP model and the RPM model are indeed special cases of the new model. We compare our estimator to Dinse's estimators through simulation and real data. Simulation study indicates that the consistent NPMLE is a good approximation to the underlying distribution for moderate sample sizes.

Entropic risk minimization for nonparametric estimation of mixing distributions

We discuss a nonparametric estimation method for the mixing distributions in mixture models. The problem is formalized as a minimization of a one-parameter objective functional, which becomes the maximum likelihood estimation or the kernel vector quantization in special cases. Generalizing the theorem for the nonparametric maximum likelihood estimation, we prove the existence and discreteness of the optimal mixing distribution and provide an algorithm to calculate it. It is demonstrated that with an appropriate choice of the parameter, the proposed method is less prone to overfitting than the maximum likelihood method. We further discuss the connection between the unifying estimation framework and the rate-distortion problem.

Model based bootstrap methods for interval censored data

The performance of model based bootstrap methods for constructing point-wise confidence intervals around the survival function with interval censored data is investigated. It is shown that bootstrapping from the nonparametric maximum likelihood estimator of the survival function is inconsistent for the current status model. A model based smoothed bootstrap procedure is proposed and proved to be consistent. In fact, a general framework for proving the consistency of any model based bootstrap scheme in the current status model is established. In addition, simulation studies are conducted to illustrate the (in)-consistency of different bootstrap methods in mixed case interval censoring. The conclusions in the interval censoring model would extend more generally to estimators in regression models that exhibit non-standard rates of convergence.

Statistical inference based on the nonparametric maximum likelihood estimator under double-truncation.

Doubly truncated data consist of samples whose observed values fall between the right- and left- truncation limits. With such samples, the distribution function of interest is estimated using the nonparametric maximum likelihood estimator (NPMLE) that is obtained through a self-consistency algorithm. Owing to the complicated asymptotic distribution of the NPMLE, the bootstrap method has been suggested for statistical inference. This paper proposes a closed-form estimator for the asymptotic covariance function of the NPMLE, which is computationally attractive alternative to bootstrapping. Furthermore, we develop various statistical inference procedures, such as confidence interval, goodness-of-fit tests, and confidence bands to demonstrate the usefulness of the proposed covariance estimator. Simulations are performed to compare the proposed method with both the bootstrap and jackknife methods. The methods are illustrated using the childhood cancer dataset.

Approximate Nonparametric Maximum Likelihood Estimation for Interval Censoring Model Case II (Running Head: NPMLE for Interval Censoring Case II)

We study the nonparametric maximum likelihood estimate of the distribution function in a type II interval censoring model. We propose an approximate solution of the problem under a technical assumption. Some basic asymptotic properties of the estimator are investigated.

Self-Consistent Nonparametric Maximum Likelihood Estimator of the Bivariate Survivor Function.

As usually formulated the nonparametric likelihood for the bivariate survivor function is over-parameterized, resulting in uniqueness problems for the corresponding nonparametric maximum likelihood estimator. Here the estimation problem is redefined to include parameters for marginal hazard rates, and for double failure hazard rates only at informative uncensored failure time grid points where there is pertinent empirical information. Double failure hazard rates at other grid points in the risk region are specified rather than estimated. With this approach the nonparametric maximum likelihood estimator is unique, and can be calculated using a two-step procedure. The first step involves setting aside all doubly censored observations that are interior to the risk region. The nonparametric maximum likelihood estimator from the remaining data turns out to be the Dabrowska (1988) estimator. The omitted doubly censored observations are included in the procedure in the second stage using self-consistency, resulting in a non-iterative nonpara-metric maximum likelihood estimator for the bivariate survivor function. Simulation evaluation and asymptotic distributional results are provided. Moderate sample size efficiency for the survivor function nonparametric maximum likelihood estimator is similar to that for the Dabrowska estimator as applied to the entire dataset, while some useful efficiency improvement arises for corresponding distribution function estimator, presumably due to the avoidance of negative mass assignments.

Goodness-of-fit tests for a semiparametric model under random double truncation

Doubly truncated data are commonly encountered in areas like medicine, astronomy, economics, among others. A semiparametric estimator of a doubly truncated random variable may be computed based on a parametric specification of the distribution function of the truncation times. This semiparametric estimator outperforms the nonparametric maximum likelihood estimator when the parametric information is correct, but might behave badly when the assumed parametric model is far off. In this paper we introduce several goodness-of-fit tests for the parametric model. The proposed tests are investigated through simulations. For illustration purposes, the tests are also applied to data on the induction time to acquired immune deficiency syndrome for blood transfusion patients.

Convex Optimization, Shape Constraints, Compound Decisions, and Empirical Bayes Rules

Estimation of mixture densities for the classical Gaussian compound decision problem and their associated (empirical) Bayes rules is considered from two new perspectives. The first, motivated by Brown and Greenshtein, introduces a nonparametric maximum likelihood estimator of the mixture density subject to a monotonicity constraint on the resulting Bayes rule. The second, motivated by Jiang and Zhang, proposes a new approach to computing the Kiefer–Wolfowitz nonparametric maximum likelihood estimator for mixtures. In contrast to prior methods for these problems, our new approaches are cast as convex optimization problems that can be efficiently solved by modern interior point methods. In particular, we show that the reformulation of the Kiefer–Wolfowitz estimator as a convex optimization problem reduces the computational effort by several orders of magnitude for typical problems, by comparison to prior EM-algorithm based methods, and thus greatly expands the practical applicability of the resulting methods. Our new procedures are compared with several existing empirical Bayes methods in simulations employing the well-established design of Johnstone and Silverman. Some further comparisons are made based on prediction of baseball batting averages. A Bernoulli mixture application is briefly considered in the penultimate section.

Nonparametric estimation of the mixing distribution in logistic regression mixed models with random intercepts and slopes

An algorithm that computes nonparametric maximum likelihood estimates of a mixing distribution for a logistic regression model containing random intercepts and slopes is proposed. The algorithm ide...

Quantifying unrecognised replication present in reports of HIV diagnoses

New diagnoses of HIV infection were reported confidentially to the Public Health Laboratory Service AIDS Centre under a national voluntary surveillance scheme. Two sets of data drawn from the national data sets were made available to us for analysis, the first in 1991 and the second in 1994, by which time the replication of reports had been reduced. The data used in the analyses consisted of the numbers of replications of the reported full date of birth in the individual records (one, two, three and so on), for each year of birth. This paper uses a nonparametric maximum likelihood estimation method for quantifying the amount of replication in the data. The estimated amount of replication was 3.37% (95% confidence interval (0.98%, 11.83%)) in the 1991 data set and 0.58% (95% confidence interval (0%, 2.64%)) in the 1994 data set.

Computing confidence intervals for log-concave densities

In Balabdaoui, Rufibach, and Wellner (2009), pointwise asymptotic theory was developed for the nonparametric maximum likelihood estimator of a log-concave density. Here, the practical aspects of their results are explored. Namely, the theory is used to develop pointwise confidence intervals for the true log-concave density. To do this, the quantiles of the limiting process are estimated and various ways of estimating the nuisance parameter appearing in the limit are studied. The finite sample size behavior of these estimated confidence intervals is then studied via a simulation study of the empirical coverage probabilities.

A Bezier curve method in interval-censored data

In this paper we propose a Bezier curve method to estimate the survival function and the median survival time in interval-censored data. We compare the proposed estimator with other existing methods such as the parametric method, the single point imputation method, and the nonparametric maximum likelihood estimator through extensive numerical studies, and it is shown that the proposed estimator performs better than others in the sense of mean squared error and mean integrated squared error. An illustrative example based on a real data set is given.

Semiparametric transformation models with random effects for clustered doubly censored data

For right-censored data, Zeng et al. [Semiparametirc transformation modes with random effects for clustered data. Statist Sin. 2008;18:355–377] proposed a class of semiparametric transformation models with random effects to formulate the effects of possibly time-dependent covariates on clustered failure times. In this article, we demonstrate that the approach of Zeng et al. can be extended to analyse clustered doubly censored data. The asymptotic properties of the nonparametric maximum likelihood estimators of the model parameters are derived. A simulation study is conducted to investigate the performance of the proposed estimators.

Convex Optimization inR

Convex optimization now plays an essential role in many facets of statistics. We briefly survey some recent developments and describe some implementations of these methods in R . Applications of linear and quadratic programming are introduced including quantile regression, the Huber M-estimator and various penalized regression methods. Applications to additively separable convex problems subject to linear equality and inequality constraints such as nonparametric density estimation and maximum likelihood estimation of general nonparametric mixture models are described, as are several cone programming problems. We focus throughout primarily on implementations in the R environment that rely on solution methods linked to R, like MOSEK by the package Rmosek. Code is provided in R to illustrate several of these problems. Other applications are available in the R package REBayes, dealing with empirical Bayes estimation of nonparametric mixture models.

Semiparametric odds rate model for modeling short-term and long-term effects with application to a breast cancer genetic study.

The proportional odds model is commonly used in the analysis of failure time data. The assumption of constant odds ratios over time in the proportional odds model, however, can be violated in some applications. Motivated by a genetic study with breast cancer patients, we propose a novel semiparametric odds rate model for the analysis of right-censored survival data. The proposed model incorporates the short-term and long-term covariate effects on the failure time data and includes the proportional odds model as a nested model. We develop efficient likelihood-based inference procedures and establish the large sample properties of the proposed nonparametric maximum likelihood estimators. Simulation studies demonstrate that the proposed methods perform well in practical settings. An application to the motivating example is provided.

Maximum-likelihood estimation of a log-concave density based on censored data

We consider nonparametric maximum-likelihood estimation of a log-concave density in case of interval-censored, right-censored and binned data. We allow for the possibility of a subprobability density with an additional mass at $+\infty$, which is estimated simultaneously. The existence of the estimator is proved under mild conditions and various theoretical aspects are given, such as certain shape and consistency properties. An EM algorithm is proposed for the approximate computation of the estimator and its performance is illustrated in two examples.