Search

The paper presents an original homogenization method to predict the elastic properties of multiphase pre-impregnated composite materials like Sheet- and Bulk Molding Compounds. The upper and lower limits of the homogenized coefficients for a 27% fibres volume fraction SMC are computed. It is presented a comparison between the upper and lower limits of the homogenized elastic coefficients of a SMC material and the experimental data. The estimation model used as a homogenization method of these heterogeneous composite materials, gave emphasis to a good agreement between this theoretical approach and experimental data. [ABSTRACT FROM AUTHOR]

This paper provides estimation of reliability of a component subjected to life testing and the procedure includes essentially polling of two samples of failure data, where the component follows exponential failure model. We assumed that some prior information is available in the form of an initial guess value (θ0) about the value of the parameter (θ) of exponential distribution from the past, and proposed a two stage shrinkage pooling estimator (TSPE) of reliability function of exponential distribution using complete failure data. The expressions for the bias, mean squared error, expected sample size and relative efficiency are derived. Conclusions regarding the constants involved in the proposed estimator are presented. Simulation, comparisons, and numerical results are reported. The proposed estimator fairs better than the classical two stage pooling estimator. [ABSTRACT FROM AUTHOR]

The treatment of incomplete data is an important step in pre-processing data prior to later analysis. We propose a novel non-parametric multiple imputation algorithm for estimating missing value. The proposed algorithm is based on Generalized Regression Neural Networks. We compare the proposed algorithm against existing algorithms on forty-five real and synthetic datasets. The effectiveness of imputation algorithms is evaluated in classification problems. The performance of proposed algorithm appears to be superior to that of other algorithms. [ABSTRACT FROM AUTHOR]

The existing methods for analyzing unreplicated factorials that do not have contemplate the possibility of outliers in experimental data have a poor performance for detecting the active effects when that possibility becomes a reality. We propose an interactive procedure based on robust regression which has a good performance in the presence or absence of contaminated data. [ABSTRACT FROM AUTHOR]

In this paper we explore image segmentation using active contours model to detect oil spills. A partial differential equation based level set method, which represents the spill surface as an implicit propagation interface, is used. Starting from an initial estimation with priori information, the level set method creates a set of speed functions to detect the position of the propagation interface. Specifically, the image intensity gradient and the curvature are utilized together to determine the speed and direction of the propagation. This allows the front interface to propagate naturally with topological changes, significant protrusions and narrow regions, giving rise to stable and smooth boundaries that discriminate oil spills from the surrounding water. The proposed method has been illustrated by experiments to detect oil spills in real images. Its advantages over the traditional image segmentation approaches have also been demonstrated. [ABSTRACT FROM AUTHOR]

This paper presents an alternative energy-based algorithm to provide speech/silence classification. The algorithm is capable to track non-stationary signals and dynamically calculate instantaneous value for threshold using adaptive scaling parameter. It is based on the observation of a noise power estimation used for computation of the threshold can be obtained using minimum and maximum values of a short-term energy estimate. The paper presents this novel voice activity detection algorithm, its performance, its limitations, and some other techniques which deal with energy estimation as well. [ABSTRACT FROM AUTHOR]

The prediction of software defects has recently attracted the attention of software quality researchers. Many predictive classification systems have already been proposed, which aim at early discovery of software modules that are fault-prone and versa. The proposed methods are usually assessed using datasets available from NASA Metrics Data repository. These datasets include a combination of design-level and codelevel metrics for different modules. To apply a defect predictor, all metrics have to be measured for any of the modules (to be used as the classifier inputs). The measurement of some of these metrics is easy and can be done straight forward. However, there are a number of metrics which are more difficult or time-consuming to quantify. Moreover, many of them do not have an exact value; so, they may get different values when using different formulas or tools. In this paper, we first discuss this hypothesis that some strong dependencies exist among various features of these datasets. Based on this hypothesis, we search for short combinations of features from the first category (easy-to-measure features), which can describe any of the features from the second category (hard-to-measure features) with a high accuracy. Then, we introduce a set of fuzzy modeling systems, each of which estimates the value of one of the second category features from its specified determinants. The evaluation of the estimation systems is carried out by computing the MSE values for all features. The experimental results are promising. The presented estimation system provides usability of the defect prediction system rather than its accuracy. Using this system, the user will not have to measure all the required mentioned metrics for any of the modules. All the features of the second category will automatically be estimated with a high accuracy. [ABSTRACT FROM AUTHOR]

Performances of speaker verification systems are superb in clean noise-free conditions but the reliability of the systems drop severely in noisy environments. Fusion of audio and visual information is one of the solutions to this limitation. However, this approach requires appropriate weighting for each biometric trait when the systems are implemented under inconsistent conditions. In this study, we propose a novel approach by introducing Support Vector Machine (SVM) as indicator system for audio reliability estimation. This approach directly validate the quality of the incoming (claimant) speech signal so as to adaptively change the weighting factor for fusion of both subsystems scores. It is important to priory check the speech signal quality because unreliable speech data give incorrect scores hence affect the accuracy of the total scores of the fusion systems. The effectiveness of this approach has been experimented to a multibiometric verification system that employs lipreading images as visual features. This verification system uses SVM as a classifier for both subsystems. Principle Component Analysis (PCA) technique is executed for visual features extraction while for the audio feature extraction; Linear Predictive Coding (LPC) technique has been utilized. In this study, we found that the SVM indicator system is able to determine the quality of the speech signal up to 99.66%. We then observed that by using the proposed adaptive fusion system, EER percentage in noisy condition (10dB) has been decreased to 0.27% compared to 9.3% for non-adaptive fusion system and 51.13% for audio only system. [ABSTRACT FROM AUTHOR]

The focus of this paper is the problem of recursive estimation for uncertain multisensor linear discrete-time systems. We herein propose a new suboptimal filtering algorithm. The basis of the proposed algorithm is the fusion formula for an arbitrary number of local Kalman filters. The proposed suboptimal filter fuses each local Kalman filter by weighted sum with scalar weights. This filter can be implemented in real time because the scalar weights do not depend on current observations in distinction to the optimal adaptive filter. The examples given, demonstrate the effectiveness and high precision of proposed filter. [ABSTRACT FROM AUTHOR]

In this paper I use Monte Carlo simulated option data to investigated the empirical power of six Risk Neutral Density (RND) estimation techniques. Three alternative approaches are used for comparison and the choice of the most suitable method depends on: their performance to correctly price options, their capacity to fit the true density - which is estimated from the undrelying assets series - and their ability to forecast the future realization of the underlying asset. I found that the decision depends on the purpose of the framework: for pricing purpose, the interpolation techniques would be adequate. However, if the aim is to extract market expectations, the Edge worth expansion should be used. [ABSTRACT FROM AUTHOR]

The objective of this paper is to study how algorithms of optimization affect the parametersestimation of Autoregressive AR(1)Models. In our research we have represented the AR(1) models in linear state space form and applied the Kalman Filters to estimate the different unknown parameters of the model. Many methods have been proposed by researchers for the estimation of the parameters in the case of the linear state space models. In our work we have emphasized on the estimation through the Maximum Likelihood (ML). Statisticians have used many algorithms to optimise the likelihood function and they have proposed many filters; publishing their results in many papers. In spite of the fact that this field is so extended, we have emphasized our study in the financial field. Two quasi-Newton algorithms: Berndt, Hall, Hall, and Hausman (BHHH) and Broyden-Fletcher-Goldfarb-Shanno (BFGS), and the Expectation-Maximization (EM) algorithm have been chosen for this study. A practical study of these algorithms applied to the maximization of likelihood by means of the Kalman Filter have been done. The results are focused on efficiency in time of computation and precision of the unknown parameters estimation. A simulation study has been carried out, using as true values the parameters of this model published in the literature, in order to test the efficiency and precision of our implemented algorithms. [ABSTRACT FROM AUTHOR]

In This paper we propose a simple and easy-to-use method for computing accurate estimate (in closed form) of the double barrier hitting time distribution of a mean-reverting lognormal process, and discuss its application to pricing exotic options whose payoffs are contingent upon barrier hitting times. This new approach is also to provide tight upper and lower bounds (in closed form) of the exact result. Within the multi-stage approximation scheme, the estimate and bounds can be easily improved in a systematic manner. Furthermore, this approach can be straight-forwardly extended to those cases with specified moving boundaries as well. [ABSTRACT FROM AUTHOR]

Experimental data need to be assessed for purposes of model identification, estimation of model parameters and consequences of misspecified model fits. Here the first and third factors are considered via analytic formulations for the distribution of the maximum likelihood estimates. When estimating this distribution with statistics, it is a tradition to invert the roles of population quantities and quantities that have been estimated from the observed sample. If the model is known, simulations, normal approximations and p*-formula methods can be used. However, exact analytic methods for describing the estimator density are recommended. One of the methods (TED) can be used when the data generating model differs from the estimation model, which allows for the estimation of common parameters across a suite of candidate models. Information criteria such as AIC can be used to pick a winning model. AIC is however approximate and generally only asymptotically correct. For fairly simple models, where expressions remain tractable, the exact estimator density under TED allows for comparisons between models. This is illustrated via a novel information criterion. Three linear models are compared and fitted to econometric data on patent filings. [ABSTRACT FROM AUTHOR]

The spatio-temporal modelling and forecasting of incidences of crimes have now become a routine part of crime prevention operations. However, obtaining reliable forecasts for cases of changing spatial boundaries using models which take the influences of exogenous factors on the spatial and temporal dynamics of crime into account remains a challenge. The proposed Dynamic Spatial Distribution Approach (DSDA) is a modelling approach that provides spatio-temporal forecasts that incorporate the influences of salient weather conditions that have a bearing on crime dynamics at both the temporal and spatial levels. The DSDA starts with modelling and forecasting the dataset of the entire area of interest as a whole. Given reliable forecasts, the DSDA spatially disaggregates them by extrapolating the spatial patterns identified from the historic data. This is carried out by firstly identifying the spatial pattern and then defining a set of weights, one for each of the clusters making up this pattern. A number of methods to determining these weights, including a benchmark, univariate and causal approaches are proposed. The DSDA was developed within a GIS and classical statistics environment. Applied to the daily number of criminal damage incidences in the City of Cardiff, UK, the DSDA yielded a city-wide forecast error of 3.51 crimes per day. The spatial disaggregation using weights identified using Least Squares estimates of expected crime percentages in each cluster yielded an average of 1.1240 crimes per day per cluster. The model used incorporated the influences of weekdays and ambient conditions on the spatial distribution of crime. [ABSTRACT FROM AUTHOR]

This paper considers parameter estimation for nonlinear and non-Gaussian state-space models with correlation. We propose an asymptotic quasilikelihood (AQL) approach which utilises a nonparametric kernel estimator of the conditional variance covariances matrix St to replace the true St in the standard quasi-likelihood. The kernel estimation avoids the risk of potential miss-specification of St and thus make the parameter estimator more robust. This has been further verified by empirical studies carried out in this paper. [ABSTRACT FROM AUTHOR]

The problem of estimating the intensity process of a doubly stochastic Poisson process is analyzed. Using covariance information, a recursive linear minimum mean-square error estimate is designed. Moreover, an efficient procedure for the computation of its associated error covariance is shown. The proposed solution becomes an alternative approach to the Kalman filter which is applicable under the only structural assumption that the intensity process to be estimated has a finite-dimensional covariance function. [ABSTRACT FROM AUTHOR]

In this paper, we present an interesting filtering algorithm to perform accurate estimation in jump Markov nonlinear systems, in case of multi-target tracking. With this paper, we aim to contribute in solving the problem of model based body motion estimation by using data coming from visual sensors. The Interacting Multiple Model (IMM) algorithm is specially designed to track accurately targets whose state and/or measurement (assumed to be linear) models changes during motion transition. However, when these models are nonlinear, the IMM algorithm must be modified in order to guarantee an accurate track. In order to deal with this problem, the IMM algorithm was combined with the Unscented Kalman Filter (UKF) [6]. Even if the later algorithm proved its efficacy in nonlinear model case; it presents a serious drawback in case of non Gaussian noise. To deal with this problem we propose to substitute the UKF with the Particle Filter (PF). To overcome the problem of data association, we propose the use of the JPDA approach, [12]. To reduce the computational burden of the latter technique, we choose firstly the most likely feasible events by applying a Genetic Algorithm; finally the derived algorithm from the combination of the IMM-PF algorithm and the GA-JPDA approach is noted GA-JPDA-IMM-PF. [ABSTRACT FROM AUTHOR]

Chapter 12 of the book "Guide to Producing National Health Accounts" is presented. It focuses on the development of the financing agents by health providers table. The first step of the process involves the breaking down of each financing agent's expenditures by provider type. It suggests the use of other estimation methods if there is no direct information on the breakdown. Table formats used by providers to report income and expenses are presented.

Positive stable laws have become a standard tool in modelling heavy tailed data in such diverse areas as finance, engineering and survival analysis. Due to the non-existence of closed-form expression for the corresponding densities, standard procedures for estimation of the parameters of positive stable distributions appear to be computationally expensive. In this note we show that the first two moments of negative order provide a straightforward estimation procedure, in which the solution of the resulting equations exists, and leads to unique moment estimates for the parameters involved. Simulations and application of this method on real data are also included. [ABSTRACT FROM AUTHOR]

The article offers information about national HIV-prevalence estimates. The method is based on surveillance data that assess prevalence over time among pregnant women who attend selected sentinel antenatal clinics. If these are disrupted during conflict and post-conflict situations, population-based surveys can be used. However, it may underestimate infection levels of participation is too low.