Excess Error Research Articles

MAP-FCRNN: Multi-step ahead prediction model using forecasting correction and RNN model with memory functions

Currently, prediction stands as one of the most prominent areas of research. Enhancing the accuracy and generalization capabilities of prediction models remains a crucial and ongoing challenge. Furthermore, the majority of existing prediction models suffer from the issue of error accumulation. Thus, we develop a multi-step time series prediction model that relies on prediction correction to address this problem. First, we mitigate the problem of excessive accumulation error by constructing a sample set in combination with the prediction target. Second, we employ a recurrent neural network (RNN) model with memory function to make initial predictions. Finally, building on the concept of prediction correction, we develop a new prediction model that effectively rectifies the initial prediction results. Remarkably, the model efficiently safeguards against deviations during prediction tasks. Additionally, our proposed model integrates a clustering algorithm during the data processing phase, which introduces a rule for sample selection. This rule ensures the inclusion of diverse types of data to enhance the prediction accuracy of the model. Notably, we conduct a comparative experimental analysis using eight publicly available data sets and evaluate our model against seven commonly used prediction models to demonstrate its effectiveness.

Study on The Method for Predicting The Explosion Source Point Based on The Characteristics of Shock Wave

To improve the accuracy in predicting the blast initiation position in tests from the measured data, a method of computing the blast source coordinates based on the average velocity of the blast shock and the corresponding correction and iteration methodology is proposed, and the accuracy of the method is verified and validated by comparing with the available data. The results show that: the accuracy of the method is good and its reliability is also high, and a spherical charge data correction method is proposed based on the characteristics of the shock wave, which can effectively improve the prediction accuracy. Using it, the coordinate error of the explosion source point is reduced from 0.656m to 0.084m, and the maximum relative error is reduced from 4.95% to 0.31%. The coordinate error of the dynamic explosion example is 1.68m, and the maximum relative error is 16.1%. The study found that the error of the original data has a great influence on prediction accuracy, and to ensure the high accuracy of the prediction results, the data with excessive errors should be excluded.

Stability-Based Generalization Analysis of the Asynchronous Decentralized SGD

The generalization ability often determines the success of machine learning algorithms in practice. Therefore, it is of great theoretical and practical importance to understand and bound the generalization error of machine learning algorithms. In this paper, we provide the first generalization results of the popular stochastic gradient descent (SGD) algorithm in the distributed asynchronous decentralized setting. Our analysis is based on the uniform stability tool, where stable means that the learned model does not change much in small variations of the training set. Under some mild assumptions, we perform a comprehensive generalizability analysis of the asynchronous decentralized SGD, including generalization error and excess generalization error bounds for the strongly convex, convex, and non-convex cases. Our theoretical results reveal the effects of the learning rate, training data size, training iterations, decentralized communication topology, and asynchronous delay on the generalization performance of the asynchronous decentralized SGD. We also study the optimization error regarding the objective function values and investigate how the initial point affects the excess generalization error. Finally, we conduct extensive experiments on MNIST, CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets to validate the theoretical findings.

A Super-resolution Algorithm for Remote Sensing Images based on Data Enhancement and Generative Adversarial Networks

In this paper, a new algorithm is proposed to address the problem of excessive errors in current super-resolution algorithms for remote sensing images, which optimizes the processing effect of generative adversarial networks by increasing the type and size of training data with data enhancement techniques. First, the algorithm constructs a more reasonable degradation model for the problem of lacking data sets of paired images, and randomly mixes and washes the degraded prior knowledge (such as blur, noise, downsampling, etc.) in the imaging process to simulate the generation process of low-resolution remote sensing images in natural scenes, and generates realistic low-resolution images for training; meanwhile, the algorithm uses ResNet34 and CNN as the basis to enhance the texture details of remote sensing images, thus richer features can be extracted from the images. The results of experiments conducted on the Alsat2 B real dataset show that the method reduces the time and cost required for sample data acquisition, optimizes the processing speed, enhances the temporal and spatial resolution of remote sensing images, and improves the effect of super-resolution processing.

Uncertainty analysis of structural output with closed-form expression based on surrogate model

Uncertainty analysis (UA) is the process that quantitatively identifies and characterizes the output uncertainty and has a crucial implication in engineering applications. The research of efficient estimation of structural output moments in probability space plays an important part in the UA and has great engineering significance. Given this point, a new UA method based on the Kriging surrogate model related to closed-form expressions for the perception of the estimation of mean and variance is proposed in this paper. The new proposed method is proven effective because of its direct reflection on the prediction uncertainty of the output moments of metamodel to quantify the accuracy level. The estimation can be completed by directly using the redefined closed-form expressions of the model’s output mean and variance to avoid excess post-processing computational costs and errors. Furthermore, a novel framework of adaptive Kriging estimating mean (AKEM) is demonstrated for more efficiently reducing uncertainty in the estimation of output moment. In the adaptive strategy of AKEM, a new learning function based on the closed-form expression is proposed. Based on the closed-form expression which modifies the computational error caused by the metamodeling uncertainty, the proposed learning function enables the updating of metamodel to reduce prediction uncertainty efficiently and realize the decrease in computational costs. Several applications are introduced to prove the effectiveness and efficiency of the AKEM compared with a universal adaptive Kriging method. Through the good performance of AKEM, its potential in engineering applications can be spotted.

A hybrid parallelization approach based on workers grouping algorithm

As the volume of model data increases, traditional machine learning is not able to train models efficiently, so distributed machine learning is gradually used in large-scale data training. Currently, commonly used distributed machine learning algorithms are based on data parallelism, and often use an overall synchronous parallel strategy when passing data, but using this strategy makes the overall training speed limited by the computation speed of the slower workers in the cluster. While the asynchronous parallel strategy maximizes the computational speed of the cluster, there is a delay in updating the parameters of the global model, which may lead to excessive computational errors or non-convergence of the model. In this paper, the author combines these two data delivery methods by grouping workers together and using synchronous parallelism for the workers in the group and asynchronous parallelism for the components for training. The experiment shows that the hybrid parallelism strategy can reduce the training time with guaranteed correctness.

Robust PCA for high‐dimensional data based on characteristic transformation

SummaryIn this paper, we propose a novel robust principal component analysis (PCA) for high‐dimensional data in the presence of various heterogeneities, in particular strong tailing and outliers. A transformation motivated by the characteristic function is constructed to improve the robustness of the classical PCA. The suggested method has the distinct advantage of dealing with heavy‐tail‐distributed data, whose covariances may be non‐existent (positively infinite, for instance), in addition to the usual outliers. The proposed approach is also a case of kernel principal component analysis (KPCA) and employs the robust and non‐linear properties via a bounded and non‐linear kernel function. The merits of the new method are illustrated by some statistical properties, including the upper bound of the excess error and the behaviour of the large eigenvalues under a spiked covariance model. Additionally, using a variety of simulations, we demonstrate the benefits of our approach over the classical PCA. Finally, using data on protein expression in mice of various genotypes in a biological study, we apply the novel robust PCA to categorise the mice and find that our approach is more effective at identifying abnormal mice than the classical PCA.

Generalization Analysis of Pairwise Learning for Ranking with Deep Neural Networks.

Pairwise learning is widely employed in ranking, similarity and metric learning, area under the curb maximization, and many other learning tasks involving sample pairs. Pairwise learning with deep neural networks was considered for ranking, but enough theoretical understanding about this topic is lacking. In this letter, we apply symmetric deep neural networks to pairwise learning for ranking with a hinge loss ϕh and carry out generalization analysis for this algorithm. A key step in our analysis is to characterize a function that minimizes the risk. This motivates us to first find the minimizer of ϕh-risk and then design our two-part deep neural networks with shared weights, which induces the antisymmetric property of the networks. We present convergence rates of the approximation error in terms of function smoothness and a noise condition and give an excess generalization error bound by means of properties of the hypothesis space generated by deep neural networks. Our analysis is based on tools from U-statistics and approximation theory.

Study on stress-deformation mechanism of concrete face rockfill dam based on centrifugal model test

The traditional method of tracking the deformation of centrifugal dam model involves burying marker points on the sidewalls, which is commonly used. However, the method is inaccurate in monitoring the deformation of the model since the constraining effect of boundary effect on the model. In this study, a quasi-distributed fiber optic monitoring system was constructed using fixed-point distributed fiber Bragg grating. Particle image velocimetry, earth pressure cell, and other electromagnetic sensors were combined to examine the rationality of the face rockfill dams centrifugal model tests. Through comparative analysis, the applicability and stability of the fixed-point distributed FBG technique in the centrifugal model test were demonstrated. It is also demonstrated that traditional monitoring methods will underestimate the deformation of face rockfill dams by 2.32 % ∼ 13.55 %. This study addresses the shortcomings of existing monitoring methods with excessive measurement errors and provides a more applicable monitoring tool for developing related experimental studies.

Robust fault detection and diagnosis of primary air data sensors in the presence of atmospheric turbulence

Abstract This paper presents a fault detection and diagnosis (FDD) algorithm for various faults in the primary air data sensors (PADS) of an aircraft in the presence of external disturbances such as atmospheric turbulence. Rapid wind variations due to turbulence induce excessive error in the externally fitted air data probe measurements, which may lead to loss of control and misinterpretations by the flight crew. In adverse environmental conditions, the FDD of air data prefers robust and adaptive air data estimates that use an analytical redundancy approach with fewer computations. The proposed method considers the kinematics of the aircraft instead of the dynamics used in the state-of-the-art algorithms. The advantage of using kinematics is that it can reduce modeling errors significantly, avoiding high false alarm rates in the FDD process. For the estimation of stable and accurate air data under external disturbance, the inertial navigation system and global positioning system (INS/GPS) output are considered instead of actual air data probe or sensor measurements. The proposed algorithm uses estimates of air data using an exponentially weighted adaptive extended Kalman filter (EW-AEKF) to detect and diagnose PADS faults, which can perform well even in the presence of uncertain noise due to atmospheric turbulence experienced during flight. The simulation was carried out to validate the algorithm with flight data obtained from the X-Plane flight simulator under moderate atmospheric turbulence. The simulation experiments were carried out using the MATLAB programming platform. The results show that the proposed method achieves satisfactory FDD performance with lower root mean square error (RMSE) and computation time than traditional EKF-based algorithms.

Combined machine learning forecasting method for short-term power load based on the dynamic weight adjustment

The load of power system exhibits evident characteristics of volatility and randomness. The traditional load forecasting algorithm usually studies and trains the historical data to obtain the load model, which makes it difficult to adapt to the load dynamic change situation, and then resulting the unreasonable inaccurate prediction. In this paper, a combinatorial machine learning model is adopted to forecast short-term power load using a dynamic adjustable weight. Firstly, a combined machine learning model is constructed using three types of algorithms including the improved long and short-term neural network, bagging algorithm, and boosting regression algorithm. The weight of each algorithm is determined dynamically by the improved error function. Secondly, the dynamic error function and the optimal weight optimization algorithm are employed so as to balance the contradiction between the speed and accuracy of dynamic adjustment. For different months or different days within a month, different weight adjustment algorithms are selected for enhancement. In addition, a penalty term is introduced to improve the algorithm accuracy and the final prediction outcomes. Finally, a practical load prediction case is simulated and compared with the traditional combined prediction model with fixed weights. It is verified that the proposed model can effectively eliminate the excessive errors caused by the poor dynamic response effect. It has a good dynamic response effect and accurate prediction. The error rate is only 1.24% when the load fluctuation is significant. This study provides a novel approach to forecasting short-term power load.

Single line noise cancellation using derivative of normalized least mean square algorithm

Suppression of noise in noisy speech signal is required in many speech enhancement applications like signal recording and transmission from one place to other. In this paper a novel single line noise cancellation system is proposed using derivative of normalized least mean spare algorithm. The proposed system has two phases. The first phase is generation of secondary reference signal from incoming primary signal itself at initial silence period and pause between two words, which is essential while adaptive filter using as noise canceller. Second phase is noise cancellation using proposed modified error data normalized step size algorithm. The performance of the proposed algorithm is compared with normalized least mean square (NLMS) algorithm and original error data normalized step size (EDNSS) algorithm using standard IEEE sentence (SP23) of Noizeus data base with different types of real world noise at different level of signal to noise ratio (SNR). The output of proposed, NLMS and EDNSS algorithm are measured with output SNR, excessive mean square error (EMSE) and misadjustment (M). The results clearly illustrates that the proposed algorithm gives improved result over conventional NLMS and EDNSS algorithm. The speed of convergence is also maintained as same conventional NLMS algorithm.

Modeling, control and analysis of a novel energy-saving oriented fuel cell anode system

Fuel cell hydrogen supply sub-system needs to provide sufficient mass flow and pressure of hydrogen to meet the working demand of the stack. The hydrogen circulation pump is often used, but causes additional parasitic power of fuel cell system. In this study, a new structure of hydrogen circulation pump is proposed to reduce the parasitic power by using the pressure potential energy of hydrogen flow. The hydrogen feeding system model is established based on the solution of differential equation of state of ideal gas, mass and energy balance. Then, considering the hydrogen intake pressure and the hydrogen excess ratio of the system, a reasonable control strategy is proposed for accurate control. Simulation results show that under the fuel cell work condition and the fuel cell vehicle operation condition, pressure difference between anode and cathode and the hydrogen excess ratio can be regulated to the conditions required by the stack, and the excess ratio error and pressure difference error between anode and cathode is 3.82e-3 and 4.52e-4 Bar, respectively. The parasitic power consumption of the circulating pump is effectively reduced about two-thirds under the fuel cell vehicle operation condition. Then, the effects of anode hydrogen pressure drop, hydrogen inlet pressure and entrainment ratio of ejector on the system are studied. Reducing the anode pressure drop, increasing the ejector entrainment ratio while reducing the hydrogen inlet pressure at low power and increasing it at high power are conducive to energy-saving of the system.

A Computational Framework for Preserving Privacy and Maintaining Utility of Geographically Aggregated Data: A Stochastic Spatial Optimization Approach

Geographically aggregated data are often considered to be safe because information can be published by group as population counts rather than by individual. Identifiable information about individuals can still be disclosed when using such data, however. Conventional methods for protecting privacy, such as data swapping, often lack transparency because they do not quantify the reduction in disclosure risk. Recent methods, such as those based on differential privacy, could significantly compromise data utility by introducing excessive error. We develop a methodological framework to address the issues of privacy protection for geographically aggregated data while preserving data utility. In this framework, individuals at high risk of disclosure are moved to other locations to protect their privacy. Two spatial optimization models are developed to optimize these moves by maximizing privacy protection while maintaining data utility. The first model relocates all at-risk individuals while minimizing the error (hence maximizing the utility). The second model assumes a budget that specifies the maximum error to be introduced and maximizes the number of at-risk individuals being relocated within the error budget. Computational experiments performed on a synthetic population data set of two counties of Ohio indicate that the proposed models are effective and efficient in balancing data utility and privacy protection for real-world applications.

Concurrent Contribution of Co-contraction to Error Reduction during Dynamic Adaptation of the Wrist.

MRI-compatible robots provide a means of studying brain function involved in complex sensorimotor learning processes, such as adaptation. To properly interpret the neural correlates of behavior measured using MRI-compatible robots, it is critical to validate the measurements of motor performance obtained via such devices. Previously, we characterized adaptation of the wrist in response to a force field applied via an MRI-compatible robot, the MR-SoftWrist. Compared to arm reaching tasks, we observed lower end magnitude of adaptation, and reductions in trajectory errors beyond those explained by adaptation. Thus, we formed two hypotheses: that the observed differences were due to measurement errors of the MR-SoftWrist; or that impedance control plays a significant role in control of wrist movements during dynamic perturbations. To test both hypotheses, we performed a two-session counterbalanced crossover study. In both sessions, participants performed wrist pointing in three force field conditions (zero force, constant, random). Participants used either the MR-SoftWrist or the UDiffWrist, a non-MRI-compatible wrist robot, for task execution in session one, and the other device in session two. To measure anticipatory co-contraction associated with impedance control, we collected surface EMG of four forearm muscles. We found no significant effect of device on behavior, validating the measurements of adaptation obtained with the MR-SoftWrist. EMG measures of co-contraction explained a significant portion of the variance in excess error reduction not attributable to adaptation. These results support the hypothesis that for the wrist, impedance control significantly contributes to reductions in trajectory errors in excess of those explained by adaptation.

Calibration and testing of measurement devices at infrasound frequencies: proof from malfunctioning devices at site

With the growing prevalence of infrasound and potential for annoyance comes the need for noise assessment. Performance validation of measuring instruments is an established necessity for reliable measurement data at conventional frequencies. However, infrasound measurements are critically dependent on the integrity of the microphone. A case study is presented showing that errors in excess of 20 dB result if the microphone diaphragm is perforated, and that such a defect cannot be detected by visual examination or with a typical sound calibrator. A further laboratory study validates the findings, and a scheme is proposed for identifying when such an issue exists.

Validation of ICESat-2 Derived Data Products on Freshwater Lakes: Bathymetry, Diffuse Attenuation Coefficient for Downwelling Irradiance (Kd ), and Particulate Backscatter Coefficient (b bp)

Monitoring large bodies of water, such as the Laurentian Great Lakes in North America, can be challenging and costly. The bathymetry, the diffuse attenuation coefficient for downwelling irradiance ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K_d</i> ), and the particulate backscattering coefficient ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b_bp</i> ) are important metrics in monitoring water quality in lakes and have typically been measured in two ways: 1) via in situ sampling campaigns, which are expensive, time-consuming, and have low spatial resolution, and 2) via passive optical imagery, which can have errors in excess of 50%. Recently, ICESat-2, an active LiDAR-based satellite, has proven effective in deriving the bathymetry, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K_d</i> , and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b_bp</i> in the global oceans. However, validation of such metrics has never been done on satellite flyovers taken on the same day as in situ measurements. Likewise, studies on freshwater environments have been limited. Here, we compare in situ data sampled from Lake Michigan and Big Glen Lake between August 13th and 14th, 2021, and results derived from an ICESat-2 flyover in the same region on August 14th, 2021.We find excellent agreement between the in situ values and the satellite-derived values for all three metrics. This suggests that ICESat-2 and other future LiDAR-based satellites will be powerful tools in monitoring large freshwater lakes.

Trajectory Tracking Control of Intelligent Driving Vehicles Based on MPC and Fuzzy PID

To improve the stability and accuracy of quintic polynomial trajectory tracking, an MPC (model predictive control) and fuzzy PID (proportional‐integral‐difference)‐ based control method are proposed. A lateral tracking controller is designed by using MPC with rule‐based horizon parameters. The lateral tracking controller controls the steering angle to reduce the lateral tracking errors. A longitudinal tracking controller is designed by using a fuzzy PID. The longitudinal controller controls the motor torque and brake pressure referring to a throttle/brake calibration table to reduce the longitudinal tracking errors. By combining the two controllers, we achieve satisfactory trajectory tracking control. Relative vehicle trajectory tracking simulation is carried out under common scenarios of quintic polynomial trajectory in the Simulink/Carsim platform. The result shows that the strategy can avoid excessive trajectory tracking errors which ensures a better performance for trajectory tracking with high safety, stability, and adaptability.

Retracted] Optimization Analysis of Power Coal‐Blending Model and Its Control System Based on Intelligent Sensor Network and Genetic Algorithm

Research and optimization of the coal‐blending system can greatly improve the variety diversity and quality stability of finished coal. However, the existing research on the coal‐blending system has some defects, such as fuzzy ash content of coal‐blending warehouse products and excessive error of the coal‐blending ratio determined by artificial experience, which brings many disadvantages and difficulties to the efficient and stable production of the coal preparation plant. The genetic algorithm and intelligent sensor network are relatively advanced technologies at present. Therefore, the optimization analysis of the coal‐blending model and its control system based on the intelligent sensor network and genetic algorithm has become a research hotspot. This paper first introduces the basic theory of the genetic algorithm and intelligent sensor network and their application in coal‐blending research, then establishes an optimized dynamic coal‐blending model based on the intelligent sensor network and genetic algorithm, and analyzes its application effect. The research results show that the coal quality prediction model can be mined from coal quality data and coal‐blending data by using the genetic algorithm and ideas, and the monitoring system based on the intelligent sensor network can monitor the abnormal state anytime and anywhere. Compared with the traditional coal‐blending method, the average error is 3.33%, and the accuracy is improved by 4.82%.

Learning performance of uncentered kernel-based principal component analysis

Principal component analysis (PCA) may be the most popular dimension reduction method. In this paper, the learning scheme of kernel PCA methods is established. Moreover, for the uncentered case, we introduce the error representation, and prove the comparison theorem that the learning error can be bounded by the excess generalization error. Under the condition that the positive eigenvalues of [Formula: see text] are all single, the satisfied error bound [Formula: see text] is deduced.