Variable Error Research Articles

Assessing the Stability of Photon-Counting CT: Insights from a Two-Year Longitudinal Study.

Among the advancements in computed tomography (CT) technology, photon-counting computed tomography (PCCT) stands out as a significant innovation, providing superior spectral imaging capabilities while simultaneously reducing radiation exposure. Its long-term stability is important for clinical care, especially longitudinal studies, but is currently unknown. This study sets out to comprehensively analyze the long-term stability of a first-generation clinical PCCT scanner. Over a two-year period, from November 2021 to November 2023, we conducted weekly identical experiments utilizing the same multi-energy CT protocol. These experiments included various tissue-mimicking inserts to rigorously assess the stability of Hounsfield Units (HU) and image noise in Virtual Monochromatic Images (VMIs) and iodine density maps. Throughout this period, notable software and hardware modifications were meticulously recorded. Each week, VMIs and iodine density maps were reconstructed and analyzed to evaluate quantitative stability over time. Spectral results consistently demonstrated the quantitative stability of PCCT. VMIs exhibited stable HU values, such as variation in relative error for VMI 70 keV measuring 0.11% and 0.30% for single-source and dual-source modes, respectively. Similarly, noise levels remained stable with slight fluctuations linked to software changes for VMI 40 and 70 keV that corresponded to changes of 8 and 1 HU, respectively. Furthermore, iodine density quantification maintained stability and showed significant improvement with software and hardware changes, especially in dual-source mode with nominal errors decreasing from 1.44 to 0.03 mg/mL. This study provides the first long-term reproducibility assessment of quantitative PCCT imaging, highlighting its potential for the clinical arena. This study indicates its long-term utility in diagnostic radiology, especially for longitudinal studies.

Investigating the Impact of Data Normalization Methods on Predicting Electricity Consumption in a Building Using different Artificial Neural Network Models.

The study investigates the impact of data normalization on the prediction of electricity consumption in buildings using four multilayer Artificial Neural Networks (ANN) algorithms: Long Short-Term Memory Networks (LSTM), Levenberg-Marquardt Back-propagation (LMBP), Recurrent Neural Networks (RNN), and General Regression Neural Network (GRNN). Four data normalization approaches, Min-Max Scaling, Mean, Z-score, and Gaussian function were assessed on experimental datasets. The LSTM algorithm, when combined with Min-Max normalization, showed the most favorable predictive capabilities, with a low Coefficient of Variation of the Root Mean Square Error (CVRMSE) of 10.3 and Normalized Mean Bias Error (NMBE) of 0.6. The remaining three normalization approaches showed satisfactory concordance with empirical data, but with slight disparities in precision. The LMBP model, when using Z-score normalization, had favorable performance in forecasting electricity consumption, but the discrepancies across the models were not significant. The Recurrent Neural Network (RNN) model, when used with Gaussian normalization, exhibited the most favorable performance, with the lowest Coefficient of Variation of Root Mean Square Error (CVRMSE) at 11.8 and Normalized Mean Biased Error (NMBE) at 0.6. The Generalized Regression Neural Network (GRNN) model, trained on unprocessed data, exhibited superior performance, with the lowest Coefficient of Variation of Root Mean Square Error (CVRMSE) at 19.2 and NMBE at 1.0. In conclusion, the study highlights the significant influence of data normalization on the predictive capabilities of various ANN models, suggesting that careful use of data normalization techniques can significantly improve the accuracy of electricity consumption forecasting in buildings.

Effect of Panel Misalignment Error Distribution on the Radiation Pattern of Leighton Chajnantor Telescope’s Antenna

The Leighton dish formerly at the Caltech Submillimeter Observatory will be reassembled to become the Leighton Chajnantor Telescope (LCT). It will be required to demonstrate high surface accuracy and controlled radiation patterns at high frequency, which are largely limited by surface error due to panel misalignment on the primary reflector. Not only the surface rms error but also the error distribution has effect on the radiation pattern of the antenna. Therefore, analyzing the effect of error distributions resulting from various panel misalignments on the radiation pattern of the antenna will facilitate better compensation of surface errors during the reassembly process. To acquire the radiation pattern of LCT’s antenna with panel misalignments, we first propose a simulation method based on physical optics and the physical theory of diffraction, which offers precise results but is time-consuming, especially at high frequencies. Then, we propose a rapid computation method that combines the method of calculating the optical path difference (OPD), the OPD fitting method, and the aperture integration method for segmented reflectors. Experimental results demonstrate that the rapid computation method is highly efficient and accurate compared to the simulation method. In order to show the effect of error distributions due to two typical panel misalignments (i.e., piston and tip-tilt) on radiation patterns, experiments are conducted using the two proposed methods for various error distributions. These experiments indicate that for the same surface rms error, smaller panel errors at smaller normalized aperture radii are more conducive to achieving improved characteristics of the radiation patterns, such as reduced peak gain losses and lower sidelobe levels. Additionally, comparison experiments also reveal that variations in piston error have a greater impact on the radiation patterns than variations in tip-tilt error under the same surface rms error.

The Effect of Single-Session Stimulating Massage on the Knee Joint Position Sense in Healthy Older Adult Men: A Randomized Crossover Trial.

Proprioception is critical for older adults to maintain their balance and prevent falling. However, massage is a convenient intervention that its beneficial effect on the proprioception is suggested. This study aimed to determine whether one session of stimulating massage of the muscles around the knee joint improves position sense in older adult men. Twenty healthy older adults participated in this blind, randomized, crossover trial. The two treatment phases were massage and rest. The washout period between interventions was a 1-week interval. The massage protocol was as follows: deep effleurage, petrissage, and tapotement for 5 minutes for the anterior (tensor fascia lata, quadriceps, sartorius, and gracilis) and posterior (hamstrings) muscles of the knee (10 minutes in total). Outcome measures were absolute, constant, and variable errors (AE, CE, and VE). Participants were assessed immediately before and after the intervention by a blinded investigator. Independent t-tests were used for statistical analyses. Massage reduced absolute error (2.77°, p = 0.01). The finding of this study confirms the beneficial impact of the massage on the joint position sense in healthy older adult men.

A user-friendly method to get automated pollen analysis from environmental samples.

Automated pollen analysis is not yet efficient on environmental samples containing many pollen taxa and debris, which are typical in most pollen-based studies. Contrary to classification, detection remains overlooked although it is the first step from which errors can propagate. Here, we investigated a simple but efficient method to automate pollen detection for environmental samples, optimizing workload and performance. We applied the YOLOv5 algorithm on samples containing debris and c. 40 Mediterranean plant taxa, designed and tested several strategies for annotation, and analyzed variation in detection errors. About 5% of pollen grains were left undetected, while 5% of debris were falsely detected as pollen. Undetected pollen was mainly in poor-quality images, or of rare and irregular morphology. Pollen detection remained effective when applied to samples never seen by the algorithm, and was not improved by spending time to provide taxonomic details. Pollen detection of a single model taxon reduced annotation workload, but was only efficient for morphologically differentiated taxa. We offer guidelines to plant scientists to analyze automatically any pollen sample, providing sound criteria to apply for detection while using common and user-friendly tools. Our method contributes to enhance the efficiency and replicability of pollen-based studies.

Analysis of the temporal and spatial characteristics of lunar reconnaissance orbiter’s orbit error based on multi-coverage narrow angle camera images

ABSTRACT Photogrammetric mapping of large area of the lunar surface using high-resolution orbiter images is very challenging due to increasing spatial resolution and lack of a comprehensive understanding of orbit error characteristics. The Lunar Reconnaissance Orbiter (LRO) Narrow Angle Camera (NAC) images are currently one of the highest-resolution images of the Moon available. The study of LRO orbit errors from the perspective of mapping applications is important for high accuracy mapping of large areas using NAC images. This paper proposes a method for orbit error estimation using multi-coverage NAC images. First, available NAC images covering the study regions are orthorectified using ISIS. Then, the impact crater feature matching algorithm is used to match the multi-coverage images. Finally, estimation of the orbit error is achieved through statistical analysis of a large number of homologous points obtained from the matching. The orbit error analyses of nine representative regions on the Moon show that in most cases the refined ephemeris provided by the LOLA/GRAIL team can significantly reduce the orbit error. The orbit errors show a sawtooth variation with a period of about one year, which have been stable for a long time and tend to increase slightly in recent years. There is a slight variation of the orbit error in space, which is even less obvious when using the refined ephemeris. The general tendency is for the far side to be larger than the near side at the same latitude, and for the higher latitudes to be larger than the lower latitudes. The proposed method and analysis results provide a reference for selection and processing of multi-coverage LROC NAC images for high-resolution large-area mapping.

Improving 2–5 Qubit Quantum Phase Estimation Circuits Using Machine Learning

Quantum computing has the potential to solve problems that are currently intractable to classical computers with algorithms like Quantum Phase Estimation (QPE); however, noise significantly hinders the performance of today’s quantum computers. Machine learning has the potential to improve the performance of QPE algorithms, especially in the presence of noise. In this work, QPE circuits were simulated with varying levels of depolarizing noise to generate datasets of QPE output. In each case, the phase being estimated was generated with a phase gate, and each circuit modeled was defined by a randomly selected phase. The model accuracy, prediction speed, overfitting level and variation in accuracy with noise level was determined for 5 machine learning algorithms. These attributes were compared to the traditional method of post-processing and a 6x–36 improvement in model performance was noted, depending on the dataset. No algorithm was a clear winner when considering these 4 criteria, as the lowest-error model (neural network) was also the slowest predictor; the algorithm with the lowest overfitting and fastest prediction time (linear regression) had the highest error level and a high degree of variation of error with noise. The XGBoost ensemble algorithm was judged to be the best tradeoff between these criteria due to its error level, prediction time and low variation of error with noise. For the first time, a machine learning model was validated using a 2-qubit datapoint obtained from an IBMQ quantum computer. The best 2-qubit model predicted within 2% of the actual phase, while the traditional method possessed a 25% error.

Cylindrical gear transmission errors influenced by temperature based on thermal network

Cylindrical gear transmission errors influenced by temperature based on thermal network

Investigating the influence of clustering techniques and parameters on a hybrid PSO-driven ANFIS model for electricity prediction

The availability of reliable electrical power, which is essential for a comfortable lifestyle worldwide, requires realistic power usage projections for electric utilities and policymakers, leading to the adoption of machine learning-based modelling tools due to the limitations of traditional power usage projection approaches. However, successful modeling of power usage in neuro-fuzzy models depends on the optimal selection of hyper-parameters. Consequently, this research looked at the major impact clustering methods and hyper-parameter modifications on a particle swarm optimization (PSO)-based adaptive neuro-fuzzy inference system (ANFIS) model. The study examined two distinct clustering methods and other key hyperparameters such as the number of clusters and cluster radius, resulting in a total of 10 sub-models. The performance of the developed models was assessed using four widely recognized performance indicators: root mean square error, mean absolute percentage error (MAPE), mean absolute error (MAE), and coefficient of variation of the root mean square error (CVRMSE). Additionally, the robustness of the optimal sub-model was evaluated by comparing it with other hybrid models based on three different PSO variants. The results revealed that the combination of the ANFIS approach and PSO, specifically with two clusters, yielded the most accurate forecasting scheme with the optimal values for MAPE (7.7778%), MAE (712.6094), CVRMSE (9.5464), and RMSE (909.4998).

Interpreting Change in Disorders of Consciousness Using the Coma Recovery Scale-Revised.

The purpose of this study was to differentiate clinically meaningful improvement or deterioration from normal fluctuations in patients with disorders of consciousness (DoC) following severe brain injury. We computed indices of responsiveness for the Coma Recovery Scale-Revised (CRS-R) using data from a clinical trial of 180 participants with DoC. We used CRS-R scores from baseline (enrollment in a clinical trial) and a 4-week follow-up assessment period for these calculations. To improve precision, we transformed ordinal CRS-R total scores (0-23 points) to equal-interval measures on a 0-100 unit scale using Rasch Measurement theory. Using the 0-100 unit total Rasch measures, we calculated distribution-based 0.5 standard deviation (SD) minimal clinically important difference, minimal detectable change using 95% confidence intervals, and conditional minimal detectable change using 95% confidence intervals. The distribution-based minimal clinically important difference evaluates group-level changes, whereas the minimal detectable change values evaluate individual-level changes. The minimal clinically important difference and minimal detectable change are derived using the overall variability across total measures at baseline and 4 weeks. The conditional minimal detectable change is generated for each possible pair of CRS-R Rasch person measures and accounts for variation in standard error across the scale. We applied these indices to determine the proportions of participants who made a change beyond measurement error within each of the two subgroups, based on treatment arm (amantadine hydrochloride or placebo) or categorization of baseline Rasch person measure to states of consciousness (i.e., unresponsive wakefulness syndrome and minimally conscious state). We compared the proportion of participants in each treatment arm who made a change according to the minimal detectable change and determined whether they also changed to another state of consciousness. CRS-R indices of responsiveness (using the 0-100 transformed scale) were as follows: 0.5SD minimal clinically important difference = 9 units, minimal detectable change = 11 units, and the conditional minimal detectable change ranged from 11 to 42 units. For the amantadine and placebo groups, 70% and 58% of participants showed change beyond measurement error using the minimal detectable change, respectively. For the unresponsive wakefulness syndrome and minimally conscious state groups, 54% and 69% of participants changed beyond measurement error using the minimal detectable change, respectively. Among 115 participants (64% of the total sample) who made a change beyond measurement error, 29 participants (25%) did not change state of consciousness. CRS-R indices of responsiveness can support clinicians and researchers in discerning when behavioral changes in patients with DoC exceed measurement error. Notably, the minimal detectable change can support the detection of patients who make a "true" change within or across states of consciousness. Our findings highlight that the continued use of ordinal scores may result in incorrect inferences about the degree and relevance of a change score.

Influence of compression garments on proprioception: Asystematic review and meta-analysis.

Compression garments (CGs) are commonly used in rehabilitation and sports contexts to enhance performance and speed up recovery. Despite the growing use of CGs in recent decades, there is no unanimous consensus on their overall influence on joint proprioception. In this current meta-analysis, we aim to fill this knowledge gap by assessing the impact of CGs on joint proprioception. We conducted a literature search across seven databases and one registry. Ultimately, we included 27 studies with 671 participants. The meta-analysis revealed that wearing CGs resulted in a significant reduction in absolute error during joint position sensing (Hedges' g: -0.64, p = 0.006) as compared to no CGs. However, further analyses of variables such as constant error (p = 0.308), variable error (p = 0.541) during joint position sense tests, threshold to detect passive motion (p = 0.757), and active movement extent discrimination (p = 0.842) did not show a significant impact of CGs. The review also identified gaps in the reporting of certain outcomes, such as parameters of CGs, reporting of performance, individual-reported outcomes, and lack of placebo comparators. Consequently, this review provides guidelines for future studies that may facilitate evidence-based synthesis and ultimately contribute to a better understanding of the overall influence of CGs on joint proprioception.

Imprecise perception of hand position during early motor adaptation.

Localizing one's body parts is important for movement control and motor learning. Recent studies have shown that the precision with which people localize their hand places constraints on motor adaptation. Although these studies have assumed that hand localization remains equally precise across learning, we show that precision decreases rapidly during early motor learning. In three experiments, healthy young participants (n = 92) repeatedly adapted to a 45° visuomotor rotation for a cycle of two to four reaches, followed by a cycle of two to four reaches with veridical feedback. Participants either used an aiming strategy that fully compensated for the rotation (experiment 1), or always aimed directly at the target, so that adaptation was implicit (experiment 2). We omitted visual feedback for the last reach of each cycle, after which participants localized their unseen hand. We observed an increase in the variability of angular localization errors when subjects used a strategy to counter the visuomotor rotation (experiment 1). This decrease in precision was less pronounced in the absence of reaiming (experiment 2), and when subjects knew that they would have to localize their hand on the upcoming trial, and could thus focus on hand position (experiment 3). We propose that strategic reaiming decreases the precision of perceived hand position, possibly due to attention to vision rather than proprioception. We discuss how these dynamics in precision during early motor learning could impact on motor control and shape the interplay between implicit and strategy-based motor adaptation.NEW & NOTEWORTHY Recent studies indicate that the precision with which people localize their hand limits implicit visuomotor learning. We found that localization precision is not static, but decreases early during learning. This decrease is pronounced when people apply a reaiming strategy to compensate for a visuomotor perturbation and is partly resistant to allocation of attention to the hand. We propose that these dynamics in position sense during learning may influence how implicit and strategy-based motor adaption interact.

Does the reliability of computational models truly improve with hierarchical modeling? Some recommendations and considerations for the assessment of model parameter reliability : Reliability of computational model parameters.

Computational modeling of behavior is increasingly being adopted as a standard methodology in psychology, cognitive neuroscience, and computational psychiatry. This approach involves estimating parameters in a computational (or cognitive) model that represents the computational processes of the underlying behavior. In this approach, the reliability of the parameter estimates is an important issue. The use of hierarchical (Bayesian) approaches, which place a prior on each model parameter of the individual participants, is thought to improve the reliability of the parameters. However, the characteristics of reliability in parameter estimates, especially when individual-level priors are assumed, as in hierarchical models, have not yet been fully discussed. Furthermore, the suitability of different reliability measures for assessing parameter reliability is not thoroughly understood. In this study, we conduct a systematic examination of these issues through theoretical analysis and numerical simulations, focusing specifically on reinforcement learning models. We note that the heterogeneity in the estimation precision of individual parameters, particularly with priors, can skew reliability measures toward individuals with higher precision. We further note that there are two factors that reduce reliability, namely estimation error and intersession variation in the true parameters, and we discuss how to evaluate these factors separately. Based on the considerations of this study, we present several recommendations and cautions for assessing the reliability of the model parameters.

Error analysis of wave intercomparison based on sampling variability

Quantitative analysis of error sources in wave measurement intercomparison under real-world conditions remains challenging due to various factors such as spatial and temporal offsets in the field, diverse configurations of wave sensors and platforms, and distinct measurement algorithms. This study introduces a modified error analysis method based on sampling variability. Measurements from two co-deployed Datawell Waverider buoys at the National Marine Test Site (Weihai) in China offer an opportunity to assess the spatial variability error. The results reveal that for non-directional wave parameters, like significant wave height and mean zero-crossing period, the wave conditions in the field can be considered stationary, with random errors primarily arising from inherent sampling variability. Spatial offsets significantly impact directional wave parameters at peak frequencies, with large deviations (greater than 20°) in peak wave direction primarily attributed to bimodal distribution. For data with significant wave heights exceeding 0.5 m, random errors in the mean wave direction at the same frequency can be mainly attributed to sampling variability. The bulk directional wave parameters weighted by the energy spectrum demonstrate lower sensitivity to spatial offsets, and there is excellent agreement between the bulk mean directions when considering only the unimodal distribution data.

Maximizing Building Energy Efficiency Through Zone-Based Conditions: An Advanced Dynamic Simulation Approach

Abstract The primary objective of this research is to maximize the building energy efficiency by identifying the zone-based conditions and employing an advanced energy simulation model. It is possible to achieve this by reducing periods of excessive heating and cooling by utilizing a calibrated energy model with 15-minute measured data. Further, by regulating the energy consumption of various functional zones corresponding to the work schedules, building's energy system timetables, and the sensible temperature comfort requirements, additional energy savings can be achieved. To show the effect of dynamic simulation on building energy consumption, periodic 15-min temperature data were collected in different zones to be used both in the calibration of the energy model and in the improvement of the current energy profile. Mean bias error and cumulative variation of the root mean squared error were chosen as a performance indicator. Utilizing measured data, over-heated and over-cooled periods were defined by manually identifying zone-based indoor comfort conditions to predict improvements in overall building energy performance. The potential energy savings that can be achieved by largely eliminating over-heating periods are calculated. In addition, energy needs of different zones were considered and zone-based scheduling and zone-based comfort conditions were applied with the implementation of demand-side management. As a result, 17% energy efficiency can be achieved with an automated heating system that controls the indoor temperatures and ensures that the temperature is always kept at the desired level. As a further improvement, 32% energy efficiency can be achieved by applying zone-based scheduling and comfort conditions.

Adjustable random linear network coding (ARLNC): a solution for data transmission in dynamic IoT computational environments

In mobile computing environments, most IoT devices connected to networks experience variable error rates and possess limited bandwidth. The conventional method of retransmitting lost information during transmission, commonly used in data transmission protocols, increases transmission delay and consumes excessive bandwidth. To overcome this issue, forward error correction techniques, e.g., Random Linear Network Coding (RLNC) can be used in data transmission. The primary challenge in RLNC-based methodologies is sustaining a consistent coding ratio during data transmission, leading to notable bandwidth usage and transmission delay in dynamic network conditions. Therefore, this study proposes a new block-based RLNC strategy known as Adjustable RLNC (ARLNC), which dynamically adjusts the coding ratio and transmission window during runtime based on the estimated network error rate calculated via receiver feedback. The calculations in this approach are performed using a Galois field with the order of 256. Furthermore, we assessed ARLNC's performance by subjecting it to various error models such as Gilbert Elliott, exponential, and constant rates and compared it with the standard RLNC. The results show that dynamically adjusting the coding ratio and transmission window size based on network conditions significantly enhances network throughput and reduces total transmission delay in most scenarios. In contrast to the conventional RLNC method employing a fixed coding ratio, the presented approach has demonstrated significant enhancements, resulting in a 73% decrease in transmission delay and a 4 times augmentation in throughput. However, in dynamic computational environments, ARLNC generally incurs higher computational costs than the standard RLNC but excels in high-performance networks.

Trajectory Deviation Estimation Method for UAV-Borne Through-Wall Radar

Mini–unmanned aerial vehicles (mini-UAVs) are emerging as a promising platform for through-wall radar to sense the enclosed space in cities, especially high-rise buildings, due to their excellent maneuverability. However, due to unavoidable environmental interference such as airflow, mini-UAVs are prone to trajectory deviation thus degrading their sensing accuracy. Most of the existing approaches model the impact of trajectory deviation into a polynomial phase error on the received signal, which cannot fit the space-variant motion error well. Moreover, the large trajectory deviations of UAVs introduce the unavoidable envelope error. This article proposes an autofocusing algorithm based on the back projection (BP) image, which directly estimates the trajectory deviations between the actual and measured track. Thus, the problem of the 2D space variability of the motion error can be circumvented. The proposed method mainly consists of two steps. First, we estimate the trajectory deviation in the line-of-sight (LOS) direction by exploring the underlying linear property of the wall embedded in the BP imaging result. Then, the estimated trajectory deviation in the LOS direction is compensated for to obtain an updated BP image, followed by a Particle Swarm Optimization (PSO) approach to estimate the trajectory deviation along the track through focusing targets behind the wall. Simulations and practical experiments show that the proposed algorithm can accurately estimate the serious trajectory deviations larger than the range resolution, improving the sensing robustness of UAV-borne through-wall radar greatly.

Quantifying accuracy and precision from continuous response data in studies of spatial perception and crossmodal recalibration

The ability to detect the absolute location of sensory stimuli can be quantified with either error-based metrics derived from single-trial localization errors or regression-based metrics derived from a linear regression of localization responses on the true stimulus locations. Here we tested the agreement between these two approaches in estimating accuracy and precision in a large sample of 188 subjects who localized auditory stimuli from different azimuthal locations. A subsample of 57 subjects was subsequently exposed to audiovisual stimuli with a consistent spatial disparity before performing the sound localization test again, allowing us to additionally test which of the different metrics best assessed correlations between the amount of crossmodal spatial recalibration and baseline localization performance. First, our findings support a distinction between accuracy and precision. Localization accuracy was mainly reflected in the overall spatial bias and was moderately correlated with precision metrics. However, in our data, the variability of single-trial localization errors (variable error in error-based metrics) and the amount by which the eccentricity of target locations was overestimated (slope in regression-based metrics) were highly correlated, suggesting that intercorrelations between individual metrics need to be carefully considered in spatial perception studies. Secondly, exposure to spatially discrepant audiovisual stimuli resulted in a shift in bias toward the side of the visual stimuli (ventriloquism aftereffect) but did not affect localization precision. The size of the aftereffect shift in bias was at least partly explainable by unspecific test repetition effects, highlighting the need to account for inter-individual baseline differences in studies of spatial learning.

Application of Bayesian models in building information modeling of residential buildings

This paper lays forth a process for continuing Bayesian calibration of full-scale building energy simulation (BES) models by making use of data obtained from building information modeling (BIM) and building energy management systems (BEMS). According to a survey taken in China, it is possible that importing data from BIM and BEMS will dramatically cut down on the amount of time and effort required for the continuing calibration of BES models. After that, the continuous calibration approach that had been developed was examined with the use of a case study that was based on the actual calibration of a building. In China, Building Information Modeling (BIM) and data on the amount of electricity used each month over the course of the preceding three years were both incorporated into the case study. According to the findings, a non-continuous technique on the test dataset fared worse than a continuous Bayesian calibration method in terms of prediction accuracy and reduced uncertainty. The normalized mean biased error (NMBE) and the coefficient of variation of the root mean square error (CVRMSE) are both discussed in this research, and comparisons are drawn between the two.

The influence of variations in ocular biometric and optical parameters on differences in refractive error.

To present a paraxial method to estimate the influence of variations in ocular biometry on changes in refractive error (S) at a population level and apply this method to literature data. Error propagation was applied to two methods of eye modelling, referred to as the simple method and the matrix method. The simple method defines S as the difference between the axial power and the whole-eye power, while the matrix method uses more accurate ray transfer matrices. These methods were applied to literature data, containing the mean ocular biometry data from the SyntEyes model, as well as populations of premature infants with or without retinopathy, full-term infants, school children and healthy and diabetic adults. Applying these equations to 1000 SyntEyes showed that changes in axial length provided the most important contribution to the variations in refractive error (57%-64%), followed by lens power/gradient index power (16%-31%) and the anterior corneal radius of curvature (10%-13%). All other components of the eye contributed <4%. For young children, the largest contributions were made by variations in axial length, lens and corneal power for the simple method (67%, 23% and 8%, respectively) and by variations in axial length, gradient lens power and anterior corneal curvature for the matrix method (55%, 21% and 14%, respectively). During myopisation, the influence of variations in axial length increased from 54.5% to 73.4%, while changes in corneal power decreased from 9.82% to 6.32%. Similarly, for the other data sets, the largest contribution was related to axial length. This analysis confirms that the changes in ocular refraction were mostly associated with variations in axial length, lens and corneal power. The relative contributions of the latter two varied, depending on the particular population.