Estimation Criterion Research Articles

An adaptive residual sub-sampling algorithm for kernel interpolation based on maximum likelihood estimations

In this paper we propose an enhanced version of the residual sub-sampling method (RSM) in Driscoll and Heryudono (2007) for adaptive interpolation by radial basis functions (RBFs). More precisely, we introduce in the context of sub-sampling methods a maximum profile likelihood estimation (MPLE) criterion for the optimal selection of the RBF shape parameter. This choice is completely automatic, provides highly reliable and accurate results for any RBFs, and, unlike the original RSM, guarantees that the RBF interpolant exists uniquely. The efficacy of this new method, called MPLE-RSM, is tested by numerical experiments on some 1D and 2D benchmark target functions.

Prevalence of multimorbidity in older adults in São Paulo, Brazil: a study with ISA-Capital.

ABSTRACTOBJECTIVETo estimate the prevalence of multimorbidity in older adults in São Paulo, Brazil.METHODSA cross-sectional study based on the 2015 ISA-Capital population-based survey, with a subsample of 1,019 older adults aged ≥ 60 years old. Multimorbidity was categorized considering two or more chronic diseases, based on a previously defined list. The data were analyzed in univariate and multiple models with Poisson regression.RESULTSThe prevalence of multimorbidity was 40% (95%CI: 36.6–43.8), being higher in women (PR a = 1.95 [compared to men]; 95%CI: 1.58–2.40), in individuals aged ≥ 75 years old (PR a = 1.25 [compared to individuals aged ≥ 60 to 64 years old]; 95%CI: 1.01–1.60), in Black people (PR a = 1.28 [compared to White people]; 95%CI: 1.04–1.59), in high-income people (PR a = 1.27 [compared to low income]; 95%CI: 1.09–1.50) and in former smokers (PR a = 1.30 [compared to those who never smoked]; 95%CI: 1.05–1.60), and lower in smokers (PR a = 0.72 [compared to those who never smoked]; 95%CI: 1.09–1.50).CONCLUSIONThe prevalence of multimorbidity was lower than that reported in most of the reviewed studies, but consistently associated with gender, age, race/skin color, smoking habit and socioeconomic status. The standardization of conceptual and methodological criteria for estimation is a challenge to relieve problems in the planning and management of health care systems for older populations.

DWDM signals noise immunity evaluation in a nonlinear fiber-optic path

The total influence of self-phase modulation, phase cross-modulation and four-wave mixing effects and amplified spontaneous emission noise on transmitted pulsed signals noise immunity in a multichannel fiberoptic transmission system with spectral division multiplexing is investigated. The investigation is carried out under the conditions of transmitting system parameters variation: the total optical power level in optical fiber, the transmission speed and the number of spectral channels. The estimate criterion of transmitted pulsed signals noise immunity is Q-factor. The total Q-factor value dependence on the total optical power level in optical fiber for various transmission speed and the number of spectral channels is demonstrated.

Estimation of Parameters and Optimality of Second-Order Spherical Designs Using Quadratic Function Relative to Non-Spherical Face centered CCD

The study presented the estimation of parameters and optimality of second-order spherical designs using quadratic model in comparison to the non-spherical face centered CCD for varying axial distances. The designs considered were equiradial design of axial distances of 1.0 and 1.414, inscribed CCD of axial distance of 1.0 and circumscribed CCD of axial distance of 1.414, the study employed sum of square error, variance estimation, D-, A-, and T-optimality criteria as well as Grand mean of these designs for quadratic model and 1 to 10 center runs were considered. The study observed that the sum of square error of the non-spherical face centred CCD is zero (0) for radial point of n=5 with 1 centre point and this result is seen to be a misleading result, because, no process is 100%. While the sum of square error of the spherical designs with axial distance of 1.0 gave minimal sum of square errors and the spherical designs with axial distance of 1.414 gave very large sum of square error. The Grand mean of the spherical and the non-spherical designs were equal or approximately equal for radial point of n=5 for centre points 1-10 inclusive. But as the radial distance increases above 5, the Grand mean of the non-spherical CCD differs significantly from those of the spherical designs. The study suggests that the non-spherical second order design is inferior to their spherical second order design counterparts. The spherical designs with axial distance of 1.414 (equiradial and circumscribed CCD) have better D-optimality, A-optimality and T-optimality than the non-spherical face centred CCD, while, the spherical designs with axial distance of 1.0 (equiradial and inscribed) has inferior D-optimality, A-optimality and T-optimality compared to the non-spherical face centred CCD with axial distance of 1.0.

State bounding and controller design for genetic regulatory networks with multiple delays and bounded disturbances

AbstractThis article addresses the problems of state bounding and state‐feedback controller design for genetic regulatory networks with multiple delays and bounded disturbances. Multiple time‐varying discrete delays and multiple constant distributed delays are involved. A delay‐dependent sufficient condition containing several simple inequalities is first given to guarantee that the state trajectories remain inside or converge exponentially into a Cartesian product of two polytopes. Furthermore, equivalent sufficient conditions directly represented by the system parameters are derived. Based on the state bounding results, one can obtain reachable set estimation and global exponential stability criteria. By designing state‐feedback controllers, the state trajectories of the close‐loop system are limited in a given Cartesian product of two polytopes. It is worthy to stress that the proposed method does not need to construct any Lyapunov–Krasovskii functional, and it can be easily realized. The effectiveness of the obtained theoretical results is illustrated by a numerical example.

Mathematical modelling of moisture sorption isotherms of some fruits

The moisture sorption isotherms of some fruits namely pear and quince were determined at 15, 30, 45 and 60oC over a range of water activity from 0.110 to 0.920 using the standard, static-gravimetric method. The experimental sorption data were fitted with generated three-parameter sorption isotherm models on Mitrevski et al., and the referent Anderson model known in the scientific and engineering literature as Guggenheim-Anderson-de Boer (GAB) model. In order to find which models give the best results, large number of numerical experiments was performed. After that, several statistical criteria for estimation and selection of the best sorption isotherm model was used. The performed statistical analysis shows that the generated three-parameter model from Mitrevski et al. gave the best fit to the sorption data of pear and quince than the referent three-parameter Anderson model. Keywords: Sorption isotherms Pear Quince Statistical criteria

How to avoid multiple scattering in strongly scattering SANS and USANS samples

Small Angle Neutron Scattering (SANS) and Ultra Small Angle Neutron Scattering (USANS) are the only available experimental techniques to provide seamless non-destructive measurements of the geometry of the accessible and inaccessible pore structure of rocks from sub-nanopore size to the scale of macropores. They have therefore become the measurement of choice for tight reservoir rocks such as organic rich shales. A simplifying assumption in the analysis is, however, that during the path of neutrons through the sample each neutron is only scattered once. Shales are samples with a high scattering power and Multiple Scattering (MS) may occur which requires special modelling for deconvolution of the results. The approach to avoid MS is to simply reduce the sample thickness to <0.15–0.5 mm. Here, wepresent a systematic method on wavelength selection and preparation of samples to optimise extraction of microstructural data and minimise parasitic errors. Experimentally measured SAS transmission (TSAS) values are used as a practical criterion for estimation of the extent of MS.Generous beamtime allocations allowed robust testing revealing that sample thicknesses can be twice as thick as predicted using the standard protocol. Analysing thicker samples is particularly beneficial for statistically relevant characterisation of heterogeneous samples making the new protocol the method of choice for such samples.

The exponentiated half logistic skew-t distribution with GARCH-type volatility models

Most financial time series have non-normal features such as heavy tails, excess kurtosis and skewness. Financial asset returns volatility is also a significant measure in financial decisions, option pricing, risk management, and portfolio selection, so it is important to create vigorous driven conditional innovation density for GARCH-type volatility models for investigating volatility in financial time series. In this research, a new exponentiated half logistic skew-t (EHLST) innovation distribution for fitting generalized autoregressive conditional heteroskedasticity (GARCH)-type models is proposed. Structural properties of the proposed distribution such as density and distribution functions, quantile function and raw moments were derived. The maximum likelihood estimation criterion discussed is to estimate the parameters of the proposed distribution via a simulation study and real-life dataset. Application on Bitcoin digital-currency price index log-returns was carried-out to inspect the performance of the GARCH-type models with EHLST innovation density relative to five innovation densities in volatility modeling. The research results showed that the GARCH-EHLST model was selected as the best for the BTC log-returns conditional variance. More so, three forecasts performance measures were used to compare the out-of-sample performance of the models, and the GARCH-EHLST model outperformed the other models in terms of superior volatility predictive capability.

Nutritional Status and Potentially Inappropriate Medications in Elderly.

(1) Background: The association between polypharmacy and malnutrition has been investigated in several studies; however, few of these specifically deepened the relationship between potentially inappropriate medication and malnutrition. With a descriptive approach, the primary aim of our study was to evaluate the impact of the nutritional status, assessed with the Mini Nutritional Assessment (MNA), on potentially inappropriate medications (PIM), estimated 10-year survival, and the risk of adverse drug reactions in elderly patients; the secondary aim was to evaluate how the Screening Tool of Older People’s Prescriptions (STOPP), Screening Tool to Alert to Right Treatment (START), and BEERS 2019 criteria identify PIM compared to nutritional status. (2) Methods: In this study, 3091 subjects were enrolled, of whom 2748 (71.7%) were women; the median age was 80 years, with an interquartile range between 75 and 85 years of age. The subjects were assessed at the outpatient service for frail older people of the University Hospital of Cagliari. The study population was evaluated for their: MNA, Charlson Comorbidity Index, 10-year survival estimation, BEERS 2019, STOPP and START criteria, and ADR Risk scores. (3) Results: We divided the study population into three groups: MNA1 (MNA score ≥ 24), MNA2 (23.5–17), and MNA3 (<17): the severity of comorbidities, STOPP and START alerts, and BEERS 2019 criteria were significantly worse in both MNA2 and MNA3 compared to MNA1—with the exception of BEERS “non-anti-infective medications that should be avoided or have their dosage reduced with varying levels of kidney function in older adults”. Moreover, the estimated 10-year survival was significantly higher in MNA1 than in MNA2 and MNA3, and also in MNA2 compared to MNA3. Finally, the ADR risk scores were significantly lower in MNA1 than in MNA2 and MNA3. (4) Conclusions: Our study demonstrated the association between nutritional status and PIM checked with the BEERS 2019 criteria, and, for the first time, with the STOPP and START criteria.

Distributed Ellipsoidal Intersection Fusion Estimation for Multi-Sensor Complex Systems.

This paper investigates the problem of distributed ellipsoidal intersection (DEI) fusion estimation for linear time-varying multi-sensor complex systems with unknown input disturbances and measurement data transmission delays. For the problem with external unknown input disturbance signals, a non-informative prior distribution is used to model the problem. A set of independent random variables obeying Bernoulli distribution is also used to describe the situation of measurement data transmission delay caused by network channel congestion, and appropriate buffer areas are added at the link nodes to retrieve the delayed transmission data values. For multi-sensor systems with complex situations, a minimum mean square error (MMSE) local estimator is designed in a Bayesian framework based on the maximum a posteriori (MAP) estimation criterion. In order to deal with the unknown correlations among the local estimators and to select the fusion estimator with lower computational complexity, the fusion estimator is designed using ellipsoidal intersection (EI) fusion technique, and the consistency of the estimator is demonstrated. In this paper, the difference between DEI fusion and distributed covariance intersection (DCI) fusion and centralized fusion estimation is analyzed by a numerical example, and the superiority of the DEI fusion method is demonstrated.

Psychometric Methods to Evaluate Measurement and Algorithmic Bias in Automated Scoring

AbstractWith the increasing use of automated scores in operational testing settings comes the need to understand the ways in which they can yield biased and unfair results. In this paper, we provide a brief survey of some of the ways in which the predictive methods used in automated scoring can lead to biased, and thus unfair automated scores. After providing definitions of fairness from machine learning and a psychometric framework to study them, we demonstrate how modeling decisions, like omitting variables, using proxy measures or confounded variables, and even the optimization criterion in estimation can lead to biased and unfair automated scores. We then introduce two simple methods for evaluating bias, evaluate their statistical properties through simulation, and apply to an item from a large‐scale reading assessment.

A critical review of univariate non-parametric estimation of first derivatives

This paper gives new guidance for selection of univariate non-parametric derivative estimators with non-iid errors in finite samples. It is shown via an extensive set of Monte Carlo simulations that the generalized criterion of Charnigo et al. (A generalized criterion for derivative estimation. Technometrics. 2011;53:238–253.) with spline smoothing performs the best in situations with minimal noise. For increased noise, generalized cross-validation of Craven and Wahba (Smoothing noisy data with spline functions. Numer Math. 1978;31:377–403) with P spline smoothing and the improved Akaike Information Criterion of Hurvich et al. (Smoothing parameter selection in nonparametric regression using an improved Akaike information criterion. J R Stat Soc Ser B. 1998;60:271–293.) with P spline smoothing are preferred. In the class of kernel smoothing and local regression methods, the local-cubic estimator of Henderson et al. (Gradient-based smoothing parameter selection for nonparametric regression estimation. J Econom. 2015;184:233–241.) generally outperforms its competitors. An internal meta-analysis separately favours the generalized method and P spline smoothing. The empirical example given provides support for use of the local-cubic estimator.

Buffet boundary prediction using RANS-based criteria and adjoint methods

In this work a procedure is proposed for a fast estimate of the transonic buffet boundary on aerofoils and wings. This algorithm aims to be an alternative to unsteady computational fluid dynamics simulations that are prohibitively expensive for complex flow cases. The algorithm combines steady Reynolds-averaged Navier-Stokes simulations with an adjoint method. Reynolds-averaged Navier-Stokes simulations were used to determine the buffet onset by means of suitable criteria, while the adjoint method was used to compute the sensitivities of the buffet parameters to the flight conditions, and draw the buffet boundary. The procedure was tested for two-dimensional configurations with available experimental data and for two different criteria. Particular attention was paid to the efficiency of the method and its ability to reduce computational cost with respect to purely RANS criteria for buffet boundary estimation. Possible variations to the procedure were also discussed in the results section together with the effect of some key parameters of the proposed algorithm.

Topology Identification of Time-Scales Complex Networks

This paper studies a topology identification problem of complex networks with dynamics on different time scales. Using the adaptive synchronization method, some criteria for a successful estimation are obtained. In particular, by regulating the original network to synchronize with an auxiliary chaotic network, this work further explores a way to avoid the precondition of linear independence. When the adaptive controller fails to achieve the outer synchronization, an impulsive control method is used. In the end, we conclude with three numerical simulations. The results obtained in this paper generalize continuous, discrete with arbitrary time step size and mixed cases.

Correction of FAGN points velocities by results of GNSS measurements

The method of determining the velocities of the Fundamental Astronomical and Geodetic Network (FAGN) points from the coordinates time series obtained from processing GNSS measurements is considered. The criteria for estimation the quality of the solution, the issues of adequate estimation of accuracy and data control are presented. Proposals for the compilation of a priori weight matrix of measurements and velocities are considered. The importance of preliminary detection and exclusion of rough measurements is noted. The results of determining the OXTK point velocities are presented, the analysis of the results is demonstrated. It is proposed to use the decision analysis tools discussed in the article when automating the process of determining the FAGS points velocities, and the listed criteria for evaluating the results should be displayed in reports. It is recommended to publish on the RGS-center website not only the coordinates and velocities of the FAGS points, but also their average square errors.

Traction Resistance Estimation Based on Multi-Method Fusion for Distributed Drive Agricultural Vehicles

This study proposes a multi-method fusion algorithm to solve the problem of coupled traction resistance with the vehicle mass for distributed drive agricultural vehicles (DDAVs), because this coupling makes estimation measurements by agricultural vehicles difficult. Indeed, the proposed method decouples the vehicle mass and traction resistance. The vehicle mass was obtained using the recursive least square method and filtering the low-frequency parts of signals of driving force and longitudinal acceleration. After obtaining estimated vehicle mass, the dynamics method was coordinated and complemented with the kinematics method to observe the traction resistance. In the low-frequency load test, statistical performance criteria (SPCs) of the mass estimation were <inline-formula> <tex-math notation="LaTeX">${R} = 0.9985$ </tex-math></inline-formula>, root mean squared error (<i>RMSE)</i> &#x003D; 0.0551 kg, and average of prediction accuracy (<i>PA)</i> &#x003D; 98.02&#x0025;. In addition, SPCs of the traction resistance estimation were <inline-formula> <tex-math notation="LaTeX">${R}=0.9655$ </tex-math></inline-formula>, <i>RMSE</i> &#x003D; 23.0472 N, and average of <i>PA</i> &#x003D; 99.28&#x0025;. In the high-frequency load test, the maximum <i>PA</i> of the mass estimation reached 98.78&#x0025;, and SPCs of the traction resistance estimation were <inline-formula> <tex-math notation="LaTeX">${R} = 0.9371$ </tex-math></inline-formula>, <i>RMSE</i> &#x003D; 1266.3933 N, and average of <i>PA</i> &#x003D; 85.62&#x0025;. Experimental tests proved that the proposed method is robust and can accurately estimate the vehicle mass and traction resistance in real time.

Poststenotic Distal Caliber Reduction Predicts Very High-Grade Proximal Internal Carotid Artery Stenosis.

Angiography relates the residual lumen to the poststenotic distal lumen (NASCET criterion) and expresses the result in percent lumen reduction. This method is not applicable when there is a collapse of the distal lumen, as seen in severe stenosis. The purpose of this study was to evaluate whether the reduced poststenotic caliber could be an additional sonographic criterion for estimation of the degree of stenosis. We measured the caliber of the distal lumen of the extracranial internal carotid artery (ICA) in 57 patients with low-grade stenosis <50% (NASCET criterion), and in 57 patients with unilateral stenosis of ≥ 50 %. All were classified following international recommendations. The criteria used included the poststenotic velocity to discriminate moderate to high-grade stenosis (50-70%) from very high-grade (≥ 80 %) stenosis. In addition, we measured the interrater reliability of the ultrasonic measurement of the lumen in the distal ICA. In the group of patients with 50% to 70% stenosis, the distal lumen was 4.3 ± 0.6 mm. In the group with very high-grade stenosis ≥ 80%, the distal lumen was 2.6±0.5mm (p<0.0001). A distal lumen of 3.2 mm or less predicted a very high-grade stenosis with a sensitivity of 0.92 and a specificity of 0.96 (AUC 0.986; 95% CI 0.97-1.00; p=0.0001). The distal poststenotic lumen is an additional ultrasonic criterion to discriminate a very high-grade ICA stenosis from a lesser degree of ICA stenosis. It can help avoid misinterpretation due to the high variability of intrastenotic peak systolic velocities (PSV) in very high-grade ICA stenoses.

Microsimulation. A Tool for Economic Analysis

Micro simulation involves modeling the behavior of individuals and other decision units taking into account the effects of policy parameters such as tax rates, eligibility rules for benefits and subsidies and compensation rates in the social security system. The model is simulated to analyze the impact of policy changes not only on mean behavior but also on the entire distribution of target variables. Micro simulation models have thus, for instance, been used to analyze how changes in the income taxes influence the tails of the income distribution (the incidence of poverty). Micro simulation complements a more traditional economic analysis both of which have pros and cons. Micro simulation is demanding in terms of modeling effort, data requirements and computer capacity. The issues of statistical inference related to micro simulation are in principle no different from those in econometric modeling generally. In practice the large scale and complex structure of a typical micro simulation model and the shortage of good micro data raise inference issues of particular relevance for micro simulation such as the choice of estimation criteria, calibration to benchmarks and model validation. Some of these issues are discussed in this paper.

Well-Conditioning Map for Tensor Spectral Solutions for Eigenvalues in 2D Irregular Planar Domains

This study proposes a diffeomorphic map for planar irregular domains, which transforms them into a unitary circle suitable for implementing spectral solutions on a complete and orthogonal basis. The domain must be enclosed by a Jordan curve that is univocally described in polar coordinates by a C 2 class function. Because of the achieved precision, general eigenvalue problems can be solved, yielding hundreds of eigenvalues and eigenfunctions in a single shot with at least ten-digit precision. In a constructive approach, there is no need for special functions or roots. The examples focus on the Helmholtz equation, but the method applies to other eigenvalue problems such as Hamiltonians. For validation, we compare all the results with the values obtained by finite elements (FE), some with other studies, and discuss sharp error estimate criteria and the asymptotic behavior of the eigenvalues. To demonstrate the generality of the approach, we present several examples with different geometries in which solutions emerge naturally. Considering the spectral accuracy, half of the examples are unpublished solutions, and the other half agree with or expand the precision of previous studies using different approaches.

Receiver operating characteristic estimation and threshold selection criteria in three-class classification problems for clustered data.

Statistical evaluation of diagnostic tests, and, more generally, of biomarkers, is a constantly developing field, in which complexity of the assessment increases with the complexity of the design under which data are collected. One particularly prevalent type of data is clustered data, where individual units are naturally nested into clusters. In these cases, Bias can arise from omission, in the evaluation process, of cluster-level effects and/or individual covariates. Focusing on the three-class case and for continuous-valued diagnostic tests, we investigate how to exploit the clustered structure of data within a linear-mixed model approach, both when the assumption of normality holds and when it does not. We provide a method for the estimation of covariate-specific receiver operating characteristic surfaces and discuss methods for the choice of optimal thresholds, proposing three possible estimators. A proof of consistency and asymptotic normality of the proposed threshold estimators is given. All considered methods are evaluated by extensive simulation experiments. As an application, we study the use of the Lysosomal Associated Membrane Protein Family Member 5 gene expression as a biomarker to distinguish among three types of glutamatergic neurons.