Accurate Estimation Model Research Articles

Improving stability for peer-to-peer multicast overlays by active measurements

The instability of the tree-like multicast overlay caused by nodes' abrupt departures is considered as one of the major problems for peer-to-peer (P2P) multicast systems. In this paper, we present a protocol for improving the overlay's stability by actively estimating the nodes' lifetime model, and combining the nodes' lifetime information with the overlay's structural properties. We use the shifted Pareto distribution to model the nodes' lifetimes in designing our protocol. To support this model, we have measured the residual lifetimes of the nodes in a popular IPTV system named PPLive [PPLive. http://www.pplive.com], and have formally analyzed the relationships between the distribution of the nodes' lifetimes, ages and their residual lifetimes under the shifted Pareto distribution model. We evaluate the overlay construction strategies, which are essential in improving the overlay's stability in our protocol, by comparing them with a number of other strategies in simulation. The experimental results indicate that our proposed protocol could improve the overlay's stability considerably, with informative but not necessarily accurate lifetime model estimation, and with limited overhead imposed on the network as well as negligible sacrifice regarding the end-to-end service latencies for the nodes on the overlay.

Validation Case Study: Prediction of Compressible Turbulent Mixing Layer Growth Rate

A recently developed validation metric is applied to the example problem of growth-rate prediction for a turbulent, compressible mixing layer. Emphasis is placed on the validation exercise as a well-defined process, which must include quantification of numerical error, evaluation of the level of confidence in the experimental data, and estimation of model accuracy. Available data for compressible mixing layers are reexamined, resulting in a new recommended data set for the study of mixing layer growth rate. Numerical solutions to the Favre-averaged Navier-Stokes equations with various two-equation turbulence closures and compressibility models are obtained and verified through formal solution verification procedures. The validation metric results are presented for each of the models, including local error estimates for detailed evaluation of model performance and global metrics that facilitate model performance comparisons.

Robust Glottal Source Estimation Based on Joint Source-Filter Model Optimization

This paper describes a robust glottal source estimation method based on a joint source-filter separation technique. In this method, the Liljencrants–Fant (LF) model, which models the glottal flow derivative, is integrated into a time-varying ARX speech production model. These two models are estimated in a joint optimization procedure, in which a Kalman filtering process is embedded for adaptively identifying the vocal tract parameters. Since the formulated joint estimation problem is a multiparameter nonlinear optimization procedure, we separate the optimization procedure into two passes. The first pass initializes the glottal source and vocal tract models by solving a quasi-convex approximate optimization problem. Having robust initial values, the joint estimation procedure determines the accuracy of model estimation implemented with a trust-region descent optimization algorithm. Experiments with synthetic and real voice signals show that the proposed method is a robust glottal source parameter estimation method with a high degree of accuracy.

Distance-Restrained Docking of Rifampicin and Rifamycin SV to RNA Polymerase Using Systematic FRET Measurements: Developing Benchmarks of Model Quality and Reliability

We are developing distance-restrained docking strategies for modeling macromolecular complexes that combine available high-resolution structures of the components and intercomponent distance restraints derived from systematic fluorescence resonance energy transfer (FRET) measurements. In this article, we consider the problem of docking small-molecule ligands within macromolecular complexes. Using simulated FRET data, we have generated a series of benchmarks that permit estimation of model accuracy based on the quantity and quality of FRET-derived distance restraints, including the number, random error, systematic error, distance distribution, and radial distribution of FRET-derived distance restraints. We find that expected model accuracy is 10Å or better for models based on: i), ≥20 restraints with up to 15% random error and no systematic error, or ii), ≥20 restraints with up to 15% random error, up to 10% systematic error, and a symmetric radial distribution of restraints. Model accuracies can be improved to 5Å or better by increasing the number of restraints to ≥40 and/or by optimizing the distance distribution of restraints. Using experimental FRET data, we have defined the positions of the binding sites within bacterial RNA polymerase of the small-molecule inhibitors rifampicin (Rif) and rifamycin SV (Rif SV). The inferred binding sites for Rif and Rif SV were located with accuracies of, respectively, 7 and 10Å relative to the crystallographically defined binding site for Rif. These accuracies agree with expectations from the benchmark simulations and suffice to indicate that the binding sites for Rif and Rif SV are located within the RNA polymerase active-center cleft, overlapping the binding site for the RNA-DNA hybrid.

Dynamic neural network modeling for nonlinear, nonstationary machine tool thermally induced error

This paper presents a new modeling methodology for nonstationary machine tool thermal errors. The method uses the dynamic neural network model to track nonlinear time-varying machine tool errors under various thermal conditions. To accommodate the nonstationary nature of the thermo-elastic process, an Integrated Recurrent Neural Network (IRNN) is introduced to identify the nonstationarity of the thermo-elastic process with a deterministic linear trend. Experiments on spindle thermal deformation are conducted to evaluate the model performance in terms of model estimation accuracy and robustness. The comparison indicates that the IRNN performs better than other modeling methods, such as, multi-variable regression analysis (MRA), multi-layer feedforward neural network (MFN), and recurrent neural network (RNN), in terms of model robustness under a variety of working conditions.

Soil sampling strategies for spatial prediction by correlation with auxiliary maps

The paper evaluates spreading of observations in feature and geographical spaces as a key to sampling optimisation for spatial prediction by correlation with auxiliary maps. Although auxiliary data are commonly used for mapping soil variables, problems associated with the design of sampling strategies are rarely examined. When generalised least-squares estimation is used, the overall prediction error depends upon spreading of points in both feature and geographical space. Allocation of points uniformly over the feature space range proportionally to the distribution of predictor (equal range stratification, or ER design) is suggested as a prudent sampling strategy when the regression model between the soil and auxiliary variables is unknown. An existing 100-observation sample from a 50 by 50 km soil survey in central Croatia was used to illustrate these concepts. It was re-sampled to 25-point datasets using different experimental designs: ER and 2 response surface designs. The designs were compared for their performance in predicting soil organic matter from elevation (univariate example) using the overall prediction error as an evaluation criterion. The ER design gave overall prediction error similar to the minmax design, suggesting that it is a good compromise between accurate model estimation and minimisation of spatial autocorrelation of residuals. In addition, the ER design was extended to the multivariate case. Four predictors (elevation, temperature, wetness index, and NDVI) were transformed to standardised principal components. The sampling points were then assigned to the components in proportion to the variance explained by a principal component analysis and following the ER design. Since stratification of the feature space results in a large number of possible points in each cluster, the spreading in geographical space can also be maximised by selecting the best of several realisations.

Quantifying the accuracy of Hammerstein model estimation

This paper investigates the accuracy of the linear component that forms part of an overall Hammerstein model-structure estimate, and a key finding is that the process of estimating the non-linear element can have a strong effect on the associated estimate of the linear dynamics. Furthermore, this effect is not explained simply by way of considering how the input spectrum is changed by the non-linearity. Instead, it arises that the linear model-estimate variability may be dominated by a term that depends on the frequency response of the linear system itself. Amongst other things, the main results derived here have experiment design implications for Hammerstein system estimation.

Cost model development using artificial neural networks

In order for the aerospace industry to achieve success in export markets through the provision of high levels of product choice, it will need to develop and economically use many new materials and manufacturing processes. Examines how the constraints imposed by changing market trends affect the identification of “cost estimating relationships” and investigates their relative benefits and limitations in terms of their effects on the overall cost model development process. A method of establishing cost estimating relationships that appears to offer benefits to the cost modelling process is that of artificial neural networks (ANNs). Using the Taguchi method, a series of experiments have been undertaken to select an appropriate network for the “turning process”. The estimation accuracy and robustness of cost models developed using suitable ANN structures have then been examined under varying conditions in order to identify guidelines.

An Empirical Evaluation of Probability Estimation with Neural Networks

Recent interest in neural networks by researchers across a wide spectrum of disciplines has provided convincing evidence of their ability to address classification problems. In this article, we consider the issue of evaluating the predictive capability of neural networks when the output values are to be treated as probabilities. We propose the use of a variant of a chi-square statistic, based on the Hosmer-Lemeshow statistic from logistic regression, to measure the goodness-of-fit of neural network models for two-group membership problems. Through experimentation with a large real-world database, we demonstrate the application of this statistic, and examine the effects of varying the number of nodes in the hidden layer on its value. Our empirical results suggest that this statistic can be very useful in identifying significant differences in the probability estimation accuracy of neural network models.

Model structure incorporated into recursive partial realization strategies

We present strategies for choosing from various canonical identifier structures in recursive partial realization algorithms for identification of state-space models. The different model structures considered in this paper have different estimation convergence rates. The relative convergence rates of the proposed deterministic algorithms are conjectured to be consistent with our (proposed) measure of the chosen model structure parameter information content. This measure, which we call the relative structural-algorithm time constant, is determined from the Fisher information matrix. This description of relative convergence rates apparently is independent of model order, so that either over- or under-determined model parameter convergence rates can be characterized by our measure of parameter information. In the proposed stochastic identifiers, the model parameter sensitivity itself is shown to have a significant effect on the initial convergence of the model parameter estimates, while final convergence rates are again strongly connected to the measure of parameter information. Some results on the minimization of the identification criterion are examined, especially those dealing with model structure influence on location and occurrence of the criterion minima. We present a strategy for increasing model convergence rates and estimation accuracy using our relative structural-algorithm time constant. Finally, we compare our method with a recently described iterative method.

Specifying Weights in Spatial Forecasting Models: The Results of Some Experiments

The application of space-time forecasting models requires the specification of a set of place-to-place weights. The existing literature contains several recommendations for this specification. By a series of Monte Carlo simulations, various weighting rules are compared in terms of estimation of model parameters and accuracy of forecasts. It is shown that specification of the weights is important, especially when sample sizes are small and the data autocorrelated. Generally, the effective area of the weight function is more important than its distance decay.