Prediction Error Analysis Research Articles

A method for predicting the energy consumption of the main driving system of a machine tool in a machining process

The machining systems that mainly consist of machine tools are numerous and are used in a wide range of applications in industry, which usually exhibit very low energy efficiency; as a result, they have great potential for energy savings and environmental emissions reduction. To achieve such energy savings, the prediction of the energy consumption of the machining process has great significance. Also, it can provide a decision-support tool for the establishment of an energy consumption quota, the energy-saving optimization of cutting parameters, energy efficiency evaluation, and so on. Although existing researches on the energy consumption prediction of machine tools have been performed, a practical method is still lacking. Therefore, a new method for predicting the energy consumption of the main driving system of a machine tool in a machining process is proposed. First, a machining process is divided into three types of periods: start-up periods, idle periods and cutting periods. Second, the energy consumption prediction models for each type of period and the total prediction model for the machining process are established. Third, by measuring energy consumption data of the start-up and idle processes at discrete speeds, the functions of the fitted curves of the energy consumption of start-up periods and idle periods are obtained, which enables the energy consumption of the start-up period and the idle period at any different speed to be predicted. Fourth, using the cutting power calculated based on the machining parameters and the additional loss coefficients obtained based on the additional loss coefficients equation set, the energy consumption of the cutting periods can be predicted. Finally, the prediction error analysis model is constructed, and the reasons why the error is not big in the prediction are expounded. The results of a case study indicate that the method is practical and has good application prospect.

Prediction of enteric methane production from dairy cows

The main objective of this study was to elucidate the potential for prediction of enteric methane (CH4) emissions from dairy cows by using predicted rumen plus hindgut digested (fermented nutrients) and total tract digested nutrients (by using NorFor) as input variables. Twenty-one experiments (78 dietary treatments) were collected. The data-set was used to develop prediction models and to test their and extant models ability to predict enteric CH4 emissions. Models were compared based on mean squared prediction error and concordance correlation coefficient (CCC) analysis. Fermented nutrients did not predict enteric CH4 emissions adequately (CCC < 0.420). Including total digested (td) nutrients in the model [CH4 (MJ/d) = −2.13 + 1.64 tdOM (kg/d) −9.74 tdFat (kg/d) + 1.64 tdNDF (kg/d)] predicted enteric CH4 emissions more precisely (CCC = 0.733), and showed an improvement in the prediction of enteric CH4 emissions over the extant models tested.

A Comparison of NARX and BP Neural Network in Short-Term Building Cooling Load Prediction

A comparison of nonlinear autoregression with exogenous inputs (NARX) neural network and back-propagation (BP) neural network in short-term prediction of building cooling load is presented in this dissertation. Both predictive models have been applied in a group of commercial buildings and analysis of prediction errors has been highlighted. Training and testing data for both prediction models have been generated from DeST (Designers Simulation Toolkits) with climate data of Shanghai. The simulation results indicate that NARX method can achieve better accuracy and generalization ability than traditional method of BP neural network. This work provides a key support in smooth and optimizing control in air-conditioning system.

Predictive use error analysis – Development of AEA, SHERPA and PHEA to better predict, identify and present use errors

In health care, the use of technical equipment plays an integral part. To achieve a high level of patient safety, it is important to avoid use errors when handling equipment. Use errors can be mitigated by performing analyses of potential use errors during the design process. One proactive analytical method for use error analysis is Predictive Use Error Analysis (PUEA), which is a further development of the methods Action Error Analysis (AEA), Systematic Human Error Reduction and Prediction Approach (SHERPA) and Predictive Human Error Analysis (PHEA). PUEA employs a detailed process for breaking down the user's tasks into steps and then identifying and investigating potential errors of use for each step. Compared with other methods, it is significant in its use of two question levels, greater inclusion of human cognition theory and that the results of the analysis are presented in matrixes. Relevance to industryThe PUEA method is useful to the industry for evaluating existing products or serving as an evaluation tool during the design process. For instance, the PUEA method has been used to evaluate user interface designs of home-care ventilators and dialysis machines. In addition, PUEA can be applied as a final risk assessment method assessment method in the product development process.

Short-Term Wind Power Prediction and Error Analysis

The prediction accuracy of wind power is important to the power system operation. Based on BP neural network used to forecast directly and time-series method used to forecast indirectly, the output wind power prediction of 4 hours in advance was studied in this paper. Simulation results showed that the performance of direct prediction is better, and the reason for that was analyzed in the paper. Finally, error analysis of prediction was researched. Comprehensive evaluation of prediction error which contains horizontal and longitudinal error evaluation was proposed.

Application of a Method to Diagnose the Source of Performance Degradation in Mpc Systems

Model Predictive Control systems may suffer from performance degradation mainly for two reasons: (i) external unmeasured disturbances are not estimated correctly, (ii) the (linear) dynamic model used by the MPC does not match (any longer) the actual process response. In this work we present the application of a method to detect when performance is not optimal, to diagnose the source of performance degradation and to propose appropriate corrections. In the simplest situation (i), optimal performance can be restored by recomputing the estimator parameters; in the other case (ii), re-identification becomes necessary. The method is based on analysis of the prediction error, i.e. the difference between the actual measured output and the corresponding model prediction, and uses three main tools: a statistical (whiteness) test on the prediction error sequence, a subspace identification method to detect the order of the input-to-prediction error system, and a nonlinear optimization algorithm to recompute optimal estimator parameters. We illustrate the effectiveness of the method on a large-scale rigorously simulated industrial process.

Characterization of Unbound Phenytoin Concentrations in Neurointensive Care Unit Patients Using a Revised Winter-Tozer Equation

Prior studies examining the accuracy of the Winter-Tozer (WT) equation for correcting total phenytoin concentrations in critically ill patients have yielded conflicting results and are limited by small sample sizes and stringent exclusion criteria, which lessen external validity. To determine whether the traditional WT equation is appropriate in correcting total phenytoin concentrations in a large sample of patients in a neurointensive care unit (NICU) and whether a new equation may be more predictive. In a retrospective study, NICU patients with reports of a concurrent total and unbound phenytoin concentration and albumin level were analyzed. Two new predictive equations were generated using a revised WT equation and regression model of baseline and laboratory characteristics. Prediction error analysis using a 20% validation cohort was conducted on all 3 equations for comparison. A total of 140 adults were included for data analysis, with data on 80% used for derivation and 20% as validation of all equations. The mean unbound phenytoin concentration was 1.4 μg/mL, which represented a free fraction of 10%. Most samples were collected within 24 hours of NICU admission. Multivariate regression analysis demonstrated that albumin, total phenytoin concentration, sex, and creatinine clearance were predictive of measured unbound phenytoin concentrations. The traditional WT equation significantly underpredicted true unbound phenytoin concentrations, with 32.1% of patients having a prediction error of more than 50% in the validation cohort. The traditional WT equation was significantly biased in underpredicting true unbound phenytoin concentrations in neurointensive care unit patients and should not be used in this setting. Two modified equations were more accurate and precise and should be considered for use when unbound phenytoin concentrations are not readily available in an NICU population.

재성형점토의 압축특성을 이용한 자연점토의 압축지수 추정

압축지수는 점토의 압축특성을 나타내는 대표적인 물성치로 압밀시험으로 직접 결정하거나, 점토의 기본 물성으로부터 경험적 방법으로 추정한다. 그러나 자연점토의 압축지수는 지반 물성치뿐 아니라 퇴적상태에 영향을 받으므로 경험적 추정방법은 지역적 한계를 지닌다. 본 연구에서는 재성형점토의 압축특성을 이용하여 지반 물성치로부터 자연 점토의 압축지수를 추정하는 새로운 방법을 제안하였다. 부산점토의 압밀시험 결과를 통해 제안방법의 적합성을 검증하였으며, 오차원인을 분석하였다. 분석결과 제안된 방법은 경험적 추정방법보다 압축지수를 정확하게 추정하였다. 제안방법의 오차는 가정사항에 의해 발생하며, 추정오차는 <TEX>$e_{cross}/e_0$</TEX>와 명확한 반비례 관계를 나타냈다. The compression index, representing the compressibility of clay, is generally obtained from the consolidation test, or predicted by empirical correlations using soil properties. However, empirical methods have regional limitations, because the compression index is affected not only by soil properties but also by site characteristics, such as deposition conditions and stress history. In this study, a method evaluating the compression index from typical soil properties is suggested using the characteristics of reconstituted clay. By analyzing the consolidation test results of Busan clay, the suggested method is verified, and the analysis of prediction error is carried out. It is shown that the proposed method evaluates the compression index more accurately than empirical methods previously suggested. The prediction errors occur by assumption, and are inversely proportional to <TEX>$e_{cross}/e_0$</TEX>, obviously.

Analysis of noisy chaotic time series prediction error

For the noisy chaotic series prediction problem, traditional methods are quite empirical, and are lacking in the analysis of the composition of the prediction error, thereby ignoring the the difference between chaotic dynamics reconstruction and prediction model. Based on the composition of actual prediction error, the predictor bias error and input disturbance error are defined in this paper and two kinds of global forecasts, ensemble least-square method and regularization method are analysed. It is shown that the ensemble least-square method is suitable for the reconstruction of chaotic dynamics, but has a greater influence on the predictor error. On the other hand, the regularization method can improve the sensitivity of the predictor, but it can be influenced by the input perturbation error. Two simulation examples are used to demonstrate the difference between the chaotic dynamical reconstruction and the establishment of prediction model, and to compare the ensamble least-square method and the regularization method, and at the same time indicate that the actual prediction error is influenced both by the input disturbance error and by the predictor error. In practice, a balance should be stricken between the two, in order to optimize the model prediction accuracy.

Implementation/optimization of moving least squares response surfaces for approximation of hurricane/storm surge and wave responses

One of the important recent advances in the field of hurricane/storm modelling has been the development of high-fidelity numerical simulation models for reliable and accurate prediction of wave and surge responses. The computational cost associated with these models has simultaneously created an incentive for researchers to investigate surrogate modelling (i.e. metamodeling) and interpolation/regression methodologies to efficiently approximate hurricane/storm responses exploiting existing databases of high-fidelity simulations. Moving least squares (MLS) response surfaces were recently proposed as such an approximation methodology, providing the ability to efficiently describe different responses of interest (such as surge and wave heights) in a large coastal region that may involve thousands of points for which the hurricane impact needs to be estimated. This paper discusses further implementation details and focuses on optimization characteristics of this surrogate modelling approach. The approximation of different response characteristics is considered, and special attention is given to predicting the storm surge for inland locations, for which the possibility of the location remaining dry needs to be additionally addressed. The optimal selection of the basis functions for the response surface and of the parameters of the MLS character of the approximation is discussed in detail, and the impact of the number of high-fidelity simulations informing the surrogate model is also investigated. Different normalizations of the response as well as choices for the objective function for the optimization problem are considered, and their impact on the accuracy of the resultant (under these choices) surrogate model is examined. Details for implementation of the methodology for efficient coastal risk assessment are reviewed, and the influence in the analysis of the model prediction error introduced through the surrogate modelling is discussed. A case study is provided, utilizing a recently developed database of high-fidelity simulations for the Hawaiian Islands.

Toward biophysical synergy: Investigating advection along the Polar Front to identify factors influencing Alaska sablefish recruitment

In fisheries stock assessment, reliable estimation of year-class strength is often hindered by lack of data on early life history stages and limited knowledge of the underlying environmental processes influencing survival through these stages. One solution to improving these estimates of year-class strength or recruitment is to first develop regional indices representing the spatial and temporal extent of a hypothesized feature influencing a species’ recruitment. These covariates should then be integrated within a population model where a variety of model selection techniques may be conducted to test for a reduction in recruitment uncertainty. The best selected model(s) may provide insight for developing hypotheses of mechanisms influencing recruitment. Here we consider the influence of a large-scale oceanographic feature, the North Pacific Polar Front, on recruitment of Alaska sablefish (Anoplopoma fimbria). Our working hypothesis is that advection of oceanic properties along the Polar Front and associated currents plays a key role in shaping the oceanographic climate of Alaskan waters and, hence, the environment that sablefish encounter during their early life history. As a first step in this investigation, we developed time series of sea surface temperature along the Polar Front mean path. We then integrated this data into the recruitment equations of the sablefish assessment base model. Model selection was based on a multistage hypothesis testing procedure combined with cross-validation and a retrospective analysis of prediction error. The impact of the best model was expressed in terms of increased precision of recruitment estimates and proportional changes in female spawning biomass for both current estimates and in future projections. The best model suggested that colder than average wintertime sea surface temperatures in the central North Pacific represent oceanic conditions that create positive recruitment events for sablefish. The incorporation of this index in the sablefish model provided moderate reduction in unexplained recruitment variability and increased future projections of spawning biomass in the medium term. Based on this result, we developed a conceptual model of three mechanisms that in combination form an ocean domain dynamic synergy (ODDS) which influences sablefish survival through the pelagic early life history stage. Successfully incorporating environmental time series into the sablefish assessment could establish a foundation for future ecosystem-based management and allow for more informed and efficient resource allocation to stakeholders.

How Well Do Lipophilicity Parameters, MEEKC Microemulsion Capacity Factor, and Plasma Protein Binding Predict CNS Tissue Binding?

Brain fraction unbound (Fu) is critical to understanding the pharmacokinetics/dynamics of central nervous system (CNS) drugs, thus several surrogate predictors have been proposed. At present, correlations between brain Fu, microemulsion electrokinetic chromatography capacity factor (MEEKC k'), plasma Fu, octanol-water partition coefficient (clogP), and LogP at pH 7.4 (clogD(7.4) ) were compared for 94 diverse molecules, and additionally for 587 compounds. MEEKC k' was a better predictor of brain Fu (r(2) = 0.74) than calculated lipophilicity parameters (clogP r(2) = 0.51-0.54, clogD(7.4) r(2) = 0.41-0.44), but it was not superior to plasma Fu (r(2) = 0.74-0.85) as a predictor of brain Fu. MEEKC k' did not predict plasma Fu(r(2) = 0.58) as well as brain Fu, and the extent of improvement over clogP or clogD(7.4) (r(2) = 0.41-0.49) was less pronounced. Although log-log-correlation analysis supported seemingly strong prediction of brain Fu both by MEEKC k' and by plasma Fu (r(2) ≥ 0.74), analysis of prediction error estimated a 10-fold and 6.9-8.6-fold prediction interval for brain Fu estimated using MEEKC k' and plasma Fu, respectively. Therefore, MEEKC k' and plasma Fu can predict the log order of CNS tissue binding, but they cannot provide truly quantitative brain Fu predictions necessary to support in-vitro-to-in-vivo extrapolations and pharmacokinetic/dynamic data interpretation.

Analysis of prediction error of intraocular lens (IOL) power calculation formulae in children aged less than two years

Most of the published studies have compared different intraocular lens (IOL) power calculation formulae in children older than two years. Children less than two years form a small subset of most such studies. Purpose of our study is to assess the accuracy of different IOL power calculation formulae in children less than two years.

PEST controlled: responsible application of inverse techniques on UK groundwater models

Abstract The inverse modelling code PEST is a tool for model parameter estimation (calibration), sensitivity analysis and predictive error analysis (uncertainty). Despite the successful use of inverse techniques in other industries, these methods are generally not adopted in UK regional groundwater modelling. This is partly due to difficulty in achieving a good representation of the conceptual models involved. In addition, some early applications of PEST to UK models were based on monitoring data of poor hydrostratigraphic resolution, leading to questionable parameter distributions. The PEST algorithm has been tried and tested worldwide and such former deficiencies can be addressed. This paper presents examples to explain how PEST can be responsibly applied to UK groundwater models to significantly improve model calibration and provide robust sensitivity data while keeping within known bounds of the conceptual model.

Synthesis Error Analysis of NC Manufacturing and Error Prediction Modelling Based on BP Neural Network Technology

The method was proposed by use of the finite element analysis software ABAQUS and the BP neural network technology to build a synthesis error prediction model of a machining-center. Firstly the finite element model of a vertical machining center CINCINNATI ARROW750 was created by use of ABAQUS ,and the cutting force induced error was analyzed which resulted from the deformation of the machining-center’s components that was caused by the cutting force ;Secondly the geometric error of the machining-center was measured by use of the laser interometer,and the sample of synthesis error was obtained. Finally the synthesis error prediction model was obtained by use BP neural network,and through the comparison of predicted value and actual value of 25 groups of samples, the feasibility of error prediction model was verified.

Evaluation of models to predict the stoichiometry of volatile fatty acid profiles in rumen fluid of lactating Holstein cows

Volatile fatty acids (VFA), produced in the rumen by microbial fermentation, are the main energy source for ruminants. The VFA profile, particularly the nonglucogenic (acetate, Ac; butyrate, Bu) to glucogenic (propionate, Pr) VFA ratio (NGR), is associated with effects on methane production, milk composition, and energy balance. The aim of this study was to evaluate extant rumen VFA stoichiometry models for their ability to predict in vivo VFA molar proportions. The models were evaluated using an independent data set consisting of 101 treatments from 24 peer-reviewed publications with lactating Holstein cows. All publications contained a full diet description, rumen pH, and rumen VFA molar proportions. Stoichiometric models were evaluated based on root mean squared prediction error (RMSPE) and concordance correlation coefficient (CCC) analysis. Of all models evaluated, the 1998 Friggens model had the lowest RMSPE for Ac and Bu (7.2 and 20.2% of observed mean, respectively). The 2006 Bannink model had the lowest RMSPE and highest CCC for Pr (14.4% and 0.70, respectively). The 2008 Bannink model had comparable predictive performance for Pr to that of the 2006 Bannink model but a larger error due to overall bias (26.2% of MSPE). The 1982 Murphy model provided the poorest prediction of Bu, with the highest RMSPE and lowest CCC (24.6% and 0.15, respectively). The 1988 Argyle and Baldwin model had the highest CCC for Ac with an intermediate RMSPE (0.47 and 8.0%, respectively). The 2006 Sveinbjörnsson model had the highest RMSPE (13.9 and 34.0%, respectively) and lowest CCC (0.31 and 0.40, respectively) for Ac and Pr. The NGR predictions had the lowest RMSPE and highest CCC in the 2 models of Bannink, whereas the lowest predictive performance was in the 2006 Sveinbjörnsson model. It appears that the type of VFA produced is not a simple linear relationship between substrate inputs and pH as currently represented. The analysis demonstrates that most rumen VFA stoichiometric approaches explain a large part of the variation in VFA molar proportions among diets, in particular for Ac, whereas predictive power for Pr and Bu differ largely among approaches. The move toward feed evaluation systems based on animal response might necessitate an improved representation of rumen fermentation, focused on improving our understanding of VFA proportions in diets that vary from the mean.

An Analysis of Prediction Errors in Fire Temperature Rise Coefficient and Fire Duration

An Analysis of Prediction Errors in Fire Temperature Rise Coefficient and Fire Duration

Application of K-Medoids with Kd-Tree for Software Fault Prediction

Software fault prediction area is subject to problems like non availability of fault data which makes the application of supervised techniques difficult. In such cases unsupervised approaches like clustering are helpful. In this paper, K-Medoids clustering approach has been applied for software fault prediction. To overcome the inherent computational complexity of KMedoids algorithm a data structure called Kd-Tree has been used to identify data agents in the datasets. Partitioning Around Medoids is applied on these data agents and this results in a set of medoids. All the remaining data points are assigned to the nearest medoids thus obtained to get the final clusters. Software fault prediction error analysis results show that our approach outperforms all unsupervised approaches in the case of one given real dataset and gives best values for the evaluation parameters. For other real datasets, our results are comparable to other techniques. Performance evaluation of our technique with other techniques has been done. Results show that our technique reduces the total number of distance calculations drastically since the number of data agents is much less than the number of data points.

Implementation of usability analysis to detect problems in the management of kitesurfing equipment

Equipment used for high-risk sports, such as kitesurfing, needs to be efficient for the intended use and the user, and thus meet the mechanical demands and provide sufficient safety. Accidents related to kitesurfing occur, and the consequences are sometimes catastrophic. One important factor is the equipment design, which can influence the type and number of injuries due to insufficient safety systems. The aim of this study was to investigate how the design of kitesurfing equipment can affect safety issues from a usability perspective in relation to the task of preparation. A focus group of 6 subjects analyzed the task of preparing kitesurfing equipment for riding, using the evaluation methods Enhanced Cognitive Walkthrough (ECW) and Predictive Use Error Analysis (PUEA). From the evaluation, a list of plausible usability problems was identified together with proposed design guidelines. The results showed that usability problems occur during preparation. In total 35 usability problems were found, of which 11 (4 from ECW and 7 from PUEA) have a great impact on safety. These 11 problems were hard to detect for the user and could result in serious consequences for the kitesurfers during riding. The analysis resulted in some general guidelines that are possible to implement on different types of kites. To conclude, education and improved design of the equipment are essential to increase the safety of the sport.

Evaluation of enteric methane prediction equations for dairy cows used in whole farm models

AbstractThe importance of evaluating greenhouse gas (GHG) emissions from dairy cows within the whole farm setting is being realized as more important than evaluating these emissions in isolation. Current whole farm models aimed at evaluating GHG emissions make use of simple regression equations to predict enteric methane (CH4) production. The objective of the current paper is to evaluate the performance of nine CH4 prediction equations that are currently being used in whole farm GHG models. Data used to evaluate the prediction equations came from a collection of individual (IND) and treatment averaged (TRT) data. Equations were compared based on mean square prediction error (MSPE) and concordance correlation coefficient (CCC) analysis. In general, predictions were poor, with root MSPE (as a percentage of observed mean) values ranging from 20.2 to 52.5 for the IND database and from 24.0 to 38.2 for the TRT database and CCC values ranging from 0.009 to 0.493 for the IND database and from 0.000 to 0.271 for the TRT database. Overall, the equations of Moe &amp; Tyrrell and IPCC Tier II performed best on the IND dataset, and the equations of Moe &amp; Tyrrell and Kirchgeßner et al., performed best on the TRT dataset. Results show that the simple more generalized equations performed worse than those that attempted to represent important aspects of diet composition, but in general significant amounts of bias and deviation of the regression slope from unity existed for all equations. The low prediction accuracy of CH4 equations in whole farm models may introduce substantial error into inventories of GHG emissions and lead to incorrect mitigation recommendations.