Asymptotic properties for self-weighted M-estimation of MEM

For a balanced two‐way mixed model, the maximum likelihood (ML) and restricted ML (REML) estimators of the variance components were obtained and compared under the non‐negativity requirements of the variance components by Lee and Kapadia (1984). In this note, for a mixed (random blocks) incomplete block model, explicit forms for the REML estimators of variance components are obtained. They are always non‐negative and have smaller mean squared error (MSE) than the analysis of variance (AOV) estimators. The asymptotic sampling variances of the maximum likelihood (ML) estimators and the REML estimators are compared and the balanced incomplete block design (BIBD) is considered as a special case. The ML estimators are shown to have smaller asymptotic variances than the REML estimators, but a numerical result in the randomized complete block design (RCBD) demonstrated that the performances of the REML and ML estimators are not much different in the MSE sense.

The paper presents the calculation method of an asymptotic relative efficiency of the least absolute deviations estimate with respect to the maximum likelihood estimate for the parameter presented in the first order autoregressive model with a stochastic coefficient. The method is based on approximation of its asymptotic relative efficiency by Taylor series. The paper considers an example of the asymptotic relative efficiency calculation for the case where diffusion of the innovation process has an approximate Gaussian distribution (Tukey distribution). It is found out that if the assumptions of the Gaussian process are performed approximately, the maximum likelihood estimate is more efficient than the least absolute deviation estimate

Let X1,…,Xn be independent and identically distributed random variables with discrete density. This paper gives two methods for smooth estimation of pi. for each i. One method may be viewed as a generalization to the case of an infinite number of nonzero cell probabilities of the Fienberg and Holland (1970) estimator for the multinomial case. Some Monte Carlo simula-tions show that for a small sample, the proposed estimators yield smaller risks under squared error loss than the usual maximum likelihood estimator (m.l.e.). However, under mild conditions the proposed estimators are shown to have the same large sample properties (strong consistency and asymptotic normality) as the m.l.e.

Maximum likelihood estimation is used widely in classical statistics. However, except in a few cases, it does not have a closed form. Furthermore, it takes time to derive the maximum likelihood estimator (MLE) owing to the use of iterative methods such as Newton–Raphson. Nonetheless, this estimation method has several advantages, chief among them being the invariance property and asymptotic normality. Based on the first approximation to the solution of the likelihood equation, we obtain an estimator that has the same asymptotic behavior as the MLE for multivariate gamma distribution. The newly proposed estimator, denoted as , is also in closed form as long as the ‐consistent initial estimator is in the closed form. Hence, we develop some closed‐form ‐consistent estimators for multivariate gamma distribution to improve the small‐sample property. is an alternative to MLE and performs better compared to MLE in terms of computation time, especially for large datasets, and stability. For the bivariate gamma distribution, the is over 130 times faster than the MLE, and as the sample size increasing, the is over 200 times faster than the MLE. Owing to the instant calculation of the proposed estimator, it can be used in state–space modeling or real‐time processing models.

Long-term electricity load forecasting plays a vital role for utilities and planners in terms of grid development and expansion planning. An overestimate of long-term electricity load will result in substantial wasted investment on the construction of excess power facilities, while an underestimate of the future load will result in insufficient generation and inadequate demand. As a first of its kind, this research proposes the use of a multiplicative error model (MEM) in forecasting electricity load for the long-term horizon. MEM originates from the structure of autoregressive conditional heteroscedasticity (ARCH) model where conditional variance is dynamically parameterized and it multiplicatively interacts with an innovation term of time-series. Historical load data, as accessed from a United States (U.S.) regional transmission operator, and recession data, accessed from the National Bureau of Economic Research, are used in this study. The superiority of considering volatility is proven by out-of-sample forecast results as well as directional accuracy during the great economic recession of 2008. Historical volatility is used to account for implied volatility. To incorporate future volatility, backtesting of MEM is performed. Two performance indicators used to assess the proposed model are: (i) loss functions in terms of mean absolute percentage error and mean squared error (for both in-sample model fit and out-of-sample forecasts) and (ii) directional accuracy.

Strong consistency and asymptotic normality are established for the maximum partial likelihood estimate of the regression parameter in Cox's regression model. Estimates are also derived for the underlying cumulative hazard function and survival distribution. We establish the asymptotic normality of these estimates and calculate the limiting variances.

It is known that various types of location privacy attacks can be carried out using a personalized transition matrix that is learned for each target user, or a population transition matrix that is common to all target users. However, since many users disclose only a small amount of location information in their daily lives, the training data can be extremely sparse. The aim of this paper is to clarify the risk of location privacy attacks in this realistic situation. To achieve this aim, we propose a learning method that uses tensor factorization (or matrix factorization) to accurately estimate personalized transition matrices (or a population transition matrix) from a small amount of training data. To avoid the difficulty in directly factorizing the personalized transition matrices (or population transition matrix), our learning method first factorizes a transition count tensor (or matrix), whose elements are the number of transition counts that the user has made, and then normalizes counts to probabilities. We focus on a localization attack, which derives an actual location of a user at a given time instant from an obfuscated trace, and compare our learning method with the maximum likelihood (ML) estimation method in both the personalized matrix mode and the population matrix mode. The experimental results using four real data sets show that the ML estimation method performs only as well as a random guess in many cases, while our learning method significantly outperforms the ML estimation method in all of the four data sets.

Weighted least absolute deviations estimation for ARFIMA time series with finite or infinite variance

The article is concerned with parameter estimation for Gaussian Markov random fields observed with additive independent identically distributed noise. In particular, we consider maximum likelihood and maximum pseudo-likelihood estimation for the noise-free case and make adjustments to the estimators in the presence of noise. The adjusted maximum likelihood estimator is computed in O(n2) time for a square image with n pixels. The estimation method is useful when only the moments of the noise are specified or when the exact maximum likelihood estimator is difficult to compute (e.g., for certain non-Gaussian noise distributions). The adjusted maximum pseudo-likelihood estimator is straightforward to calculate, is useful as a starting value in maximization routines, and is often a reasonable estimator in its own right. We discuss asymptotic properties of the adjusted estimators including consistency. We also consider constrained maximum pseudo-likelihood estimation and a Bayesian estimator. We compare the adjusted estimators with the exact Gaussian maximum likelihood estimator and toroidal boundary approximation Gaussian maximum likelihood estimator in a simulation study. The adjusted estimators are also robust to the specification of the noise distribution.

In this paper, we propose a disturbance covariance matrix estimation method for radar signal processing when the number of training samples is severely limited. Several maximum likelihood (ML) estimation methods based on the structured of the covariance matrix, which is modeled as a sum of an unknown positive semi-definite Hermitian matrix and a matrix proportional to identity, have been proposed to solve the problem. Based on the covariance structure, we incorporate additional constraint on the low rank property of the covariance matrix, and propose a low rank regularized ML estimation method for covariance matrix estimation. The low rank is achieved by solving a reweighted trace minimization problem. Experimental results demonstrate the effective of the proposed method, both in the accuracy of rank estimation and the improvement of signal interference and noise ratio (SINR).

A maximum likelihood (ML) estimator is developed for determining time delay between signals received at two spatially separated sensors in the presence of uncorrelated noise. This ML estimator can be realized as a pair of receiver prefilters followed by a cross correlator. The time argument at which the correlator achieves a maximum is the delay estimate. The ML estimator is compared with several other proposed processors of similar form. Under certain conditions the ML estimator is shown to be identical to one proposed by Hannan and Thomson [10] and MacDonald and Schultheiss [21]. Qualitatively, the role of the prefilters is to accentuate the signal passed to the correlator at frequencies for which the signal-to-noise (S/N) ratio is highest and, simultaneously, to suppress the noise power. The same type of prefiltering is provided by the generalized Eckart filter, which maximizes the S/N ratio of the correlator output. For low S/N ratio, the ML estimator is shown to be equivalent to Eckart prefiltering.

This study investigates the performance of four parameter estimation methods for the Laplace distribution: Method of Moments (MM), Maximum Likelihood Estimation (MLE), Minimum Chi-Square Estimation using equiprobable cells (MCE-EQ), and Minimum Chi-Square Estimation using Representative Points (MCE-RP). Through comprehensive Monte Carlo simulations with sample sizes ranging from 50 to 400, we compare the root mean squared error (RMSE) of the location (μ) and scale (b) parameter estimates. Our results demonstrate that while MLE remains robust for location estimation, the MCE-RP method consistently outperforms other estimators—including MLE—for the scale parameter, particularly in small to moderate samples. The use of Representative Points, which provide an optimal discretization of the distribution, significantly enhances estimation precision. These findings are especially relevant for medical research, where accurate estimation of variability—such as in biomarker concentration levels or physiological response times—is critical for reliable sample size determination, risk assessment, and clinical decision-making. MCE-RP thus offers a superior, reliable estimator for the Laplace scale parameter, with direct implications for improving statistical inference in applied biomedical studies. Purpose: The purpose of this research is to empirically evaluate and compare the finite-sample performance of four estimation methods for the Laplace distribution’s parameters, with a focus on the novel application of Representative Points in minimum chi-square estimation. This work seeks to bridge the gap between theoretical estimation methods and practical applications, providing applied researchers with a more robust estimation tool when modeling data with Laplace characteristics, such as those commonly encountered in medical and biomedical studies. Methods: We conducted an extensive Monte Carlo simulation study to compare the four estimation methods: MM, MLE, MCE-EQ, and MCE-RP. For each method, we generated independent and identically distributed samples from a standard Laplace distribution (μ=0, b=1) with sample sizes n = 50, 100, 200, and 400. Each scenario was replicated 1,000 times. The performance of each estimator was assessed using the root mean squared error (RMSE) for both μ and b. The MCE-RP method utilized pre-computed Representative Points for the standard Laplace distribution, which were transformed according to preliminary MLE estimates to form an optimal cell structure for chi-square minimization. All nonlinear optimizations required for MCE-EQ and MCE-RP were implemented programmatically. Results: The simulation results indicate that MLE performs best for estimating the location parameter μ across all sample sizes. However, for the scale parameter b, the MCE-RP method consistently yields lower RMSE values compared to MLE, MM, and MCE-EQ. In many cases, particularly for smaller samples, the RMSE of MCE-RP is approximately half that of MLE for b. The advantage of MCE-RP is evident across varying numbers of Representative Points (m = 5, 10, 15, 20), with optimal performance often observed at m = 10 or 15. These findings confirm that MCE-RP provides a more precise and reliable estimator for the scale parameter, making it particularly advantageous in small-sample settings. Contribution: This paper contributes to the statistical methodology for the Laplace distribution by introducing and validating the use of Representative Points within a minimum chi-square estimation framework. The key contributions are: (1) demonstrating that MCE-RP significantly outperforms established methods for estimating the scale parameter; (2) providing empirical evidence that RP-based discretization enhances estimation efficiency, especially in finite samples; (3) offering practical guidance for applied researchers in fields such as medical statistics, where accurate scale estimation is crucial for variability assessment, power analysis, and reliable inference; and (4) laying a methodological foundation for extending the RP approach to other location-scale distributions.

This paper considers a partially non linear model E(Y|X, z, t) = f(X, β) + zTg(t) and gives its T-type estimate, which is a weighted quasi-likelihood estimate using sieve method and can be obtained by EM algorithm. The influence functions and asymptotic properties of T-type estimate (consistency and asymptotic normality) are discussed, and convergence rate of both parametric and non parametric components are obtained. Simulation results show the shape of influence functions and prove that the T-type estimate performs quite well. The proposed estimate is also applied to a data set and compared with the least square estimate and least absolute deviation estimate.

A multiplicative error model with time-varying parameters and an error term following a mixture of gamma distributions is introduced. The model is fitted to the daily realized volatility series of deutschemark/dollar and yen/dollar returns and is shown to capture the conditional distribution of these variables better than the commonly used autoregressive fractionally integrated moving average model. The forecasting performance of the new model is found to be, in general, superior to that of the set of volatility models recently considered by Andersen et al. (2003, Econometrica 71, 579--625) for the same data. Copyright 2006, Oxford University Press.

This study presents a new six-DOF force/torque sensor and its calibration method. This calibration method apply the method of so-called maximum likelihood estimation (MLE). MLE is utilized to determine and identify the parameters related to applied torques/forces towards resulted deformations at varied locations of the sensor structure. Formulating such relations in a vector–matrix form, those parameters are captured as coefficients in a matrix related to torques/forces towards angular/linear deformations in different directions. In addition to applying MLE, finite element modeling and analysis are conducted to generate realistic-like empirical data for the afore-mentioned calibration computation. The matrix formed by these coefficients can predict exactly forces and torques by this sensor based on deformations detected by strain gauges. In simulated results the worst sum of error for three forces is less than 0.04% while the worst sum of error for three torques is less than 0.005%. Experiments are also conducted, and the results show that the designed sensor and maximum-likelihood parameter estimation approach achieve favorable performance of predicting forces and torques with error less than 1%. The developed sensor is suitable for real-time sensing of multi-dimensional interactive forces/torques in a robot arm.