Kernel Estimation Method Research Articles

On the Evaluation of Economic Benefits of Japanese Telemedicine and Factors for Its Promotion

This paper attempts to estimate economic values of telemedicine and to extract factors that promote telemedicine through the use of survey data. This is the first analysis aimed at evaluating telemedicine in Japan. We utilized the Contingent Valuation Method and estimated demand functions of telemedicine. Because the number of institutions replying with willingness to pay (WTP) and willingness to undertake (WTU) is relatively small, the Kernel Estimation Method was applied. After estimating WTP and WTU, by multiplying the number of medical institutions currently implementing telemedicine, the whole value in Japan was obtained. By using the Tobit Model, factors that influence WTP and WTU were extracted. Estimated WTP was 35.23 dollars for teleradiology and 162.89 dollars for telepathology. WTU estimated was 39.81 dollars for teleradiology and 86.59 dollars for telepathology. Estimated economic benefits in Japan for 1 year were 1.27 million dollars for WTP of teleradiology and 278,600 dollars for telepathology. WTU is 10 million dollars for teleradiology and 393,400 dollars for telepathology. Medical institutions with the following characteristics tend to reply larger WTP: (1) university hospitals, (2) internal medicine, (3) radiology, and (4) use of video conference systems. Regarding WTU, the following characteristics influence WTU: (1) use of a personal computer, (2) use of telepathology equipment, (3) high satisfaction with the quality of telemedicine, and (4) experience base of telemedicine. Based on the nationwide survey on telemedicine, the total value of telemedicine in Japan was estimated. In addition to the evaluation, quantitative aspects of implementation of telemedicine, such as factors to promote telemedicine, are analyzed. These can provide useful information for further implementation of telemedicine not only in Japan but in other countries as well.

An economic evaluation of Japanese telemedicine, focusing on teleradiology and telepathology

We collected information by postal survey from 622 medical institutions reported to be using telemedicine in Japan. The questionnaire asked about willingness to pay (WTP) for telemedicine and willingness to undertake (WTU) it. The Kernel Estimation Method was used to obtain WTP for teleradiology (4379 yen), telepathology (9526 yen), teleconferencing (2084 yen) and teleconsultation (633 yen). The estimated WTU were teleradiology (3875 yen), telepathology (17,918 yen), teleconferencing (3230 yen) and teleconsultation (3643 yen). These values are larger than the current charges that customers actually pay or providers receive. Multiplying these values by the number of medical institutions that had implemented telemedicine allows the total annual benefits of telemedicine to be estimated, e.g. for teleradiology, the annual benefit in terms of WTP and WTU (millions of yen per year) were 140.20 and 1101.75, respectively. Based on the results of the survey, specific policy measures to promote telemedicine further include improvement of quality and reimbursement.

Image edge detection based on multi-fractal spectrum analysis

In this paper, an image edge detection method based on multi-fractal spectrum analysis is presented. The coarse grain Holder exponent of the image pixels is first computed. then, its multi-fractal spectrum is estimated by the kernel estimation method. Finally, the image edge detection is done by means of different multi-fractal spectrum values. Simulation results show that this method is efficient and has better locality compared with the traditional edge detection methods such as the Sobel method.

Empirical Bayes test for scale exponential family

In this paper, we consider the empirical Bayes (EB) test problem for the scale parameters in the scale exponential family with a weighted linear loss function. The EB test rules are constructed by the kernel estimation method. The asymptotical optimality and convergence rates of the EB test rules are obtained. The main results are illustrated by applying the proposed test to type II censored data from the exponential distribution and to the test problem for the dispersion parameter in the linear regression model.

Semiparametric Estimation in General Repeated Measures Problems

SummaryThe paper considers a wide class of semiparametric problems with a parametric part for some covariate effects and repeated evaluations of a nonparametric function. Special cases in our approach include marginal models for longitudinal or clustered data, conditional logistic regression for matched case–control studies, multivariate measurement error models, generalized linear mixed models with a semiparametric component, and many others. We propose profile kernel and backfitting estimation methods for these problems, derive their asymptotic distributions and show that in likelihood problems the methods are semiparametric efficient. Although generally not true, it transpires that with our methods profiling and backfitting are asymptotically equivalent. We also consider pseudolikelihood methods where some nuisance parameters are estimated from a different algorithm. The methods proposed are evaluated by using simulation studies and applied to the Kenya haemoglobin data.

EMPIRICAL BAYES TEST PROBLEMS OF VARIANCE COMPONENTS IN RANDOM EFFECTS MODEL

Bayes decision rule of variance components for one-way random effects model is derived and empirical Bayes (EB) decision rules are constructed by kernel estimation method. Under suitable conditions, it is shown that the proposed EB decision rules are asymptotically optimal with convergence rates near O(n−1/2). Finally, an example concerning the main result is given.

ASSESSING THE VARIABILITY OF RISK SURFACES USING RESIDENTIAL HISTORY DATA IN A CASE CONTROL STUDY OF BREAST CANCER

ISEE-529 Introduction: Residential location has often been used as a measure of environmental exposure in epidemiologic studies. To examine breast cancer risk associated with residential history, we explored the spatio-temporal patterns of risk surfaces using data on lifetime residential history in a case control study of breast cancer. We applied GIS-based exploratory spatial analyses to obtain risk surfaces, and assessed spatio-temporal variability of the risk surfaces. Methods: A population-based case control study of breast cancer in western New York (the WEB Study) includes data on the lifetime residential history for breast cancer cases and controls. Participants were asked to provide all locations of earlier residences. Density surfaces of cases and controls were obtained using kernel estimation methods, and the standardized difference in density surfaces was identified to depict elevated areas of breast cancer risk. The significance of the resulting risk surfaces was tested and reported as p-values. These surfaces were compared for premenopausal and postmenopausal women. To assess the variabililty of risk surfaces in space and time, the standarized difference in density surfaces was obtained for specific time periods in a woman’s life and for each decade, from 1940s to 1990s. Results: We found strong evidence of clustering of lifetime residence for premenopausal women (for cases relative to controls), and little evidence of such clustering for postmenopausal women. We also identified the time points contributing most significantly to this result. When density surfaces between cases and controls were compared at each time point, we observed that the earlier decades and early time points, such as residence at birth and menarche, were more likely to be influential time points in understanding overall patterns, relative to the importance of later time points. Discussion: We were able to pinpoint geographic areas with higher risk, and to assess temporal variabililty of the risk surfaces by identifying the role of early exposures through exploratory spatial analyses.

Mapping of cellular compartments based on ultrastructural immunogold labeling

Ultrastructural identification of subcellular morphologically inconspicuous compartments is based on detection of specific molecules or by a presence of specific functions. Such compartments are detected using antibodies with attached label, usually gold particles. However, the gold particles have a point pattern, while a compartment is a coherent area. In addition, some background labeling is always present that complicates identification of the labeled compartments. The aim of this study was therefore to develop a stereological method that would enable us to define cellular compartments based on delineating the borders of gold particle clusters, and to test the practical use of the method using biological experimental data. New computer program plug-ins were developed to facilitate the practical use of the stereological method. The kernel estimation method was successfully tested by detection of ribosomal rRNA over morphologically recognizable nucleoli. In a next step, we successfully detected individual chromosomal domains—nuclear compartments that cannot be distinguished in cell nuclei morphologically. The results show that the new stereological/image analysis method is well able to discriminate cellular compartments based on density of immunogold particles. The plug-ins were made available to scientific community at http://nucleus.biomed.cas.cz/gold.

Adaptive R-Estimation in Autoregressions

In this paper we explore the local asymptotic normality property with a ranked residual central sequence in autoregression to give locally asymptotically minimax estimators of the autoregressive parameter. Then, we use the kernel estimator method described by Koul and Schick [Koul, H. L., Schick, A. (1997). Efficient estimation in nonlinear autoregressive time series models. Bernoulli 3:247–277] to construct adaptive estimators. A simulation experiment is carried out to illustrate the performance of the proposed adaptive estimators.

Adaptive R-estimation in a linear regression model with ARMA errors

This paper proposes a class of adaptive estimators based on ranks for the parameter of a linear regression model with ARMA errors. Under local asymptotic normality, these estimators are shown to be locally asymptotically minimax. Adaptivity is obtained via two distinct approaches: the first one relies on a consistent estimator of the innovation density computed from the order statistic of residuals, and the second one on an estimator of the score function by adapting the kernel estimator method.

On empirical Bayes estimation of variance components in random effects model

In this paper, Bayes estimators of variance components are derived for the one-way random effects model, and empirical Bayes (EB) estimators are constructed by the kernel estimation method of a multivariate density and its mixed partial derivatives. It is shown that the EB estimators are asymptotically optimal and convergence rates are established. Finally, an example concerning the main results is given.

Identification of Volterra kernels using interpolation

This paper presents a new method for the identification of frequency-domain Volterra kernels. Since the nonlinear kernels often play a secondary role compared to the dominant, linear component of the system, it is worth establishing a balance between the degree of liberty of these components and their effect on the overall accuracy of the model. This is necessary in order to reduce the model complexity, hence the required measurement length. Based on the assumption that frequency-domain kernels are locally smooth, the kernel surfaces can be approximated by interpolation techniques, thus reducing the complexity of the model. Similarly to the unreduced (Volterra) model, this smaller model is also i) linear in the unknowns; ii) only locally sensitive to its parameters; and iii) free of structural assumptions about the system. The parameter estimation boils down to solving a linear system of equations in the least-squares (LS) sense. The design of the interpolation scheme is described and the performance of the approximation is analyzed and illustrated by simulation. The algorithm allows a significant saving in measurement time compared to other kernel estimation methods.

USING THE KERNEL METHOD FOR FUNCTIONAL ESTIMATION WITH CURRENT STATUS DATA

Yang ([1]) proposes the use of a simple estimator of the cumulative distribution function for estimating the functionals with current status data. It is a kernel estimator and is an alternative to the nonparametric maximum likelihood estimator (NPMLE), while the resulting functional estimator has the same asymptotic normal distribution as the NPMLE based estimator. The performance of the kernel based functional estimator very much depends on the choice of bandwidth. In this article, we study the choice of bandwidth for the kernel estimator method of Yang. Through extensive numerical studies, it was found that the bandwidth can be properly chosen using the Jackknife resampling method, and that the kernel based estimator performs well when using a good choice of the bandwidth.

A probabilistic model for appearance-based robot localization

In this paper we present a method for an appearance-based modeling of the environment of a mobile robot. We describe the task (localization of the robot) in a probabilistic framework. Linear image features are extracted using a Principal Component Analysis. The appearance model is represented as a probability density function of the image feature vector given the location of the robot. We estimate this density model from the data with a kernel estimation method. We show how the parameters of the model influence the localization performance. We also study how many features and which features are needed for good localization.

A comparison of selectivity estimators for range queries on metric attributes

In this paper, we present a comparison of nonparametric estimation methods for computing approximations of the selectivities of queries, in particular range queries. In contrast to previous studies, the focus of our comparison is on metric attributes with large domains which occur for example in spatial and temporal databases. We also assume that only small sample sets of the required relations are available for estimating the selectivity. In addition to the popular histogram estimators, our comparison includes so-called kernel estimation methods. Although these methods have been proven to be among the most accurate estimators known in statistics, they have not been considered for selectivity estimation of database queries, so far. We first show how to generate kernel estimators that deliver accurate approximate selectivities of queries. Thereafter, we reveal that two parameters, the number of samples and the so-called smoothing parameter, are important for the accuracy of both kernel estimators and histogram estimators. For histogram estimators, the smoothing parameter determines the number of bins (histogram classes). We first present the optimal smoothing parameter as a function of the number of samples and show how to compute approximations of the optimal parameter. Moreover, we propose a new selectivity estimator that can be viewed as an hybrid of histogram and kernel estimators. Experimental results show the performance of different estimators in practice. We found in our experiments that kernel estimators are most efficient for continuously distributed data sets, whereas for our real data sets the hybrid technique is most promising.

Characterizing Three‐Dimensional Hydraulic Conductivity Distributions Using Qualitative and Quantitative Geologic Borehole Data: Application to a Field Site

AbstractThe focus of this paper is the characterization of three‐dimensional hydraulic conductivity distributions using qualitative and quantitative geologic borehole data. We illustrate an approach that entails a three‐step procedure, where lithologic information reported in the borehole logs is first classified into texture classes. Representative hydraulic conductivity values are then calculated for each texture class using a correlation relating measured lithologic and hydraulic data from core samples. The generalized kernel estimator method, which can take full advantage of the characteristics of borehole log data and is applicable to statistically nonstationary systems, is used thereafter to generate the three‐dimensional distributions of hydraulic conductivity. An application using borehole data from a trichloroethene (TCE)‐contaminated Superfund site in Tucson, Arizona, is presented and used to test the proposed methodology. The simulated hydraulic conductivity distributions appear reasonable and the simulated permeability distributions for several cross sections agree well with the hydrogeologic cross‐sectional maps. The use of mean permeability indicators for differentiating zones of high and low permeability is also investigated.

Comparative nonlinear modeling of renal autoregulation in rats: Volterra approach versus artificial neural networks

Volterra models have been increasingly popular in modeling studies of nonlinear physiological systems. In this paper, feedforward artificial neural networks with two types of activation functions (sigmoidal and polynomial) are utilized for modeling the nonlinear dynamic relation between renal blood pressure and flow data, and their performance is compared to Volterra models obtained by use of the leading kernel estimation method based on Laguerre expansions. The results for the two types of artificial neural networks (sigmoidal and polynomial) and the Volterra models are comparable in terms of normalized mean-square error (NMSE) of the respective output prediction for independent testing data. However, the Volterra models obtained via the Laguerre expansion technique achieve this prediction NMSE with approximately half the number of free parameters relative to either neural-network model. Nonetheless, both approaches are deemed effective in modeling nonlinear dynamic systems and their cooperative use is recommended in general, since they may exhibit different strengths and weaknesses depending on the specific characteristics of each application.

Testing independence by nonparametric kernel method

Using nonparametric kernel estimation method, we propose a consistent test for independence of two random vectors based on the L 2 norm of difference between the joint density and the product of their marginals. A Monte Carlo study is carried out to examine the finite sample performance of the proposed test.

A simple recourse model for power dispatch under uncertain demand

Optimal power dispatch under uncertainty of power demand is tackled via a stochastic programming model with simple recourse. The decision variables correspond to generation policies of a system comprising thermal units, pumped storage plants and energy contracts. The paper is a case study to test the kernel estimation method in the context of stochastic programming. Kernel estimates are used to approximate the unknown probability distribution of power demand. General stability results from stochastic programming yield the asymptotic stability of optimal solutions. Kernel estimates lead to favourable numerical properties of the recourse model (no numerical integration, the optimization problem is smooth convex and of moderate dimension). Test runs based on real-life data are reported. We compute the value of the stochastic solution for different problem instances and compare the stochastic programming solution with deterministic solutions involving adjusted demand portions.

How sensitive are average derivatives?

Average derivatives are the mean slopes of regression functions. In practice they are estimated via a nonparametric smoothing technique. Every smoothing method needs a calibration parameter that determines the finite sample performance. In this paper we use the kernel estimation method and develop a formula for the bandwidth that describes the sensitivity of the average derivative estimator. One can determine an optimal smoothing parameter from this formula which tries out to undersmooth the density of the regression variable.