Error Loss Research Articles

Reliability estimation in multicomponent stress-strength model using weighted exponential-Lindley distribution

This research investigates the reliability estimation in multicomponent stress-strength model when both the stress and strengths are drawn from Weighted Exponential-Lindley distribution. Reliability assessment is carried out using classical and Bayesian approaches. The Bayes estimates of the reliability in the multicomponent stress-strength model are derived under a squared error loss function using informative and non-informative priors for the parameters. Further, Lindley's approximation and Gibbs sampling method are used to develop Bayes estimators for the system's reliability due to the lack of explicit forms. Additionally, an asymptotic confidence interval and the highest probability density credible interval are constructed to gauge system performance. A simulation study is conducted to assess the performance of reliability estimators. Finally, a real data set is analysed for illustrative purposes.

The Efficiency of Hazard Rate Preservation Method for Generating Discrete Rayleigh–Lindley Distribution

In this study, we introduce two novel discrete counterparts for the Rayleigh–Lindley mixture, constructed through the application of survival and hazard rate preservation techniques. These two-parameter discrete models demonstrate exceptional adaptability across various data types, including skewed, symmetric, and monotonic datasets. Statistical analyses were conducted using maximum likelihood estimation and Bayesian approaches to assess these models. The Bayesian analysis, in particular, was implemented with the squared error and LINEX loss functions, incorporating a modified Lwin Prior distribution for parameter estimation. Through simulation studies and numerical methods, we evaluated the estimators’ performance and compared the effectiveness of the two discrete adaptations of the Rayleigh–Lindley distribution. The simulations reveal that Bayesian methods are especially effective in this setting due to their flexibility and adaptability. They provide more precise and dependable estimates for the discrete Rayleigh–Lindley model, especially when using the hazard rate preservation method. This method is a compelling alternative to the traditional survival discretization approach, showcasing its significant potential in enhancing model accuracy and applicability. Furthermore, two real data sets are analyzed to assess the performance of each analog.

Predicting Compressive Strength of High-Performance Concrete Using Hybridization of Nature-Inspired Metaheuristic and Gradient Boosting Machine

This study proposes a novel integration of the Extreme Gradient Boosting Machine (XGBoost) and Differential Flower Pollination (DFP) for constructing an intelligent method to predict the compressive strength (CS) of high-performance concrete (HPC) mixes. The former is employed to generalize a mapping function between the mechanical property of concrete and its influencing factors. DFP, as a metaheuristic algorithm, is employed to optimize the learning phase of XGBoost and reach a fine balance between the two goals of model building: reducing the prediction error and maximizing the generalization capability. To construct the proposed method, a historical dataset consisting of 400 samples was collected from previous studies. The model’s performance is reliably assessed via multiple experiments and Wilcoxon signed-rank tests. The hybrid DFP-XGBoost is able to achieve good predictive outcomes with a root mean square error of 5.27, a mean absolute percentage error of 6.74%, and a coefficient of determination of 0.94. Additionally, quantile regression based on XGBoost is performed to construct interval predictions of the CS of HPC. Notably, an asymmetric error loss is used to diminish overestimations committed by the model. It was found that this loss function successfully reduced the percentage of overestimated CS values from 47.1% to 27.5%. Hence, DFP-XGBoost can be a promising approach for accurately and reliably estimating the CS of untested HPC mixes.

Optimization of financial data processing strategy based on LSTM-GCN

In the process of training LSTM-GCN mixed model with labeled data, a common practice is to divide the data set into a training set and a validation set, and cross-validation techniques are frequently used to ensure the effective partitioning of data. The training phase mainly involves the selection of optimization methods such as gradient descent to update the model parameters, so that the model can correctly identify the feature patterns in the data. In this process, we first define a loss function to quantify the difference between the model's predicted value and the actual value. For example, when dealing with classification problems, the cross-entropy loss function is often used; Whereas in regression problems, the mean square error loss function or the absolute error loss function is more common. Then, the loss gradient of each parameter is calculated by backpropagation algorithm, and optimization strategies such as gradient descent are applied to adjust the parameter values. Through repeated iterations, these algorithms aim to minimize the loss function as a core goal for training the model. In addition, the details of model training, such as setting the appropriate learning rate, taking the appropriate regularization measures, and choosing the appropriate weight initialization strategy, are also key steps that cannot be ignored. This paper conforms to the research and development trend of the industry, focusing on how to use the LSTM-GCN model to capture temporal and spatial characteristics, process multi-modal data, process nonlinear data and dynamic weight allocation, etc., providing new ideas for Internet finance companies in data processing. This model combines the advantages of LSTM and GCN, and has prominent advantages such as timeliness, strong representation learning ability, associability, and interpretability. It can process both time series data and graphic data, and can analyze and mine the correlation and regularity in large-scale time series data, so as to process financial data more accurately. It greatly reduces the risk of investment decision.

Reliability Estimation for Burr XII Distribution under the Weighted Q-Symmetric Entropy Loss Function

Considering that the choice of loss function plays a significant role in the derivation of Bayesian estimators, we propose a novel asymmetric loss function named the weighted Q-symmetric entropy loss for computing the estimates of the parameter and reliability function of the Burr XII distribution. This paper covers the classical maximum-likelihood, uniformly minimum-variance unbiased, and Bayesian estimation methods under the squared error loss, general entropy loss, Q-symmetric entropy loss, and new loss functions. Through Monte Carlo simulation, the respective performances of the considered estimators for the reliability function are evaluated, indicating that the Bayesian estimator under the new loss function is more efficient than those under other loss functions. Finally, a real data set is used to demonstrate the practicality of the presented estimators.

Bayesian inference for the inverse Weibull distribution based on symmetric and asymmetric balanced loss functions with application

In this study, the unified hybrid censored approach is employed to estimate the parameters of the inverse Weibull distribution, as well as the survival and hazard rate functions. Parameter estimates are obtained using both Bayesian and Maximum Likelihood approaches, with Bayesian estimates acquired through Lindley's approximation method using three distinct balanced loss functions. These encompass both symmetric and asymmetric balanced loss functions, specifically the balanced squared error (BSE) loss function, the balanced linear exponential (BLINEX) loss function, and the balanced general entropy (BGE) loss function. We conduct a simulation study to compare the effectiveness of various estimators, and a real-world data analysis is presented to illustrate practical implementation. Ultimately, our findings indicate that Bayesian parameter estimates consistently outperform their Maximum Likelihood counterparts across all methods.

Design and Experiment of Plate Taking Control System of Edible Sunflower (Edulis helianthus catino L.) Harvester

This study aims at the problems of high labor intensity, high cost and high loss rate of mechanical picking of seeds, low mechanization level and difficulty to guarantee the quality of picking plates in the process of picking edible sunflower. Based on the principle of manual plate taking, a plate taking control system for a sunflower harvesting table was designed. The principle of taking the plate of the edible sunflower harvesting table was analyzed. According to the actual operation requirements, the overall scheme of the sunflower plate control system is determined, and the control strategy of imitating artificial low-loss harvesting is designed. To reduce the grain loss in the process of taking the plate and improve the control accuracy of the system on the movement speed and displacement of the key components in the process of taking the plate, a trapezoidal acceleration and deceleration control algorithm is designed as the control algorithm of taking the plate. The working performance of the plate control system was verified with the absolute error, relative error and total loss rate of the harvest as objectives. Bench and field verification experiments were both carried out. The bench experiment showed that the speed error of the plate parts was not more than 0.028 m/s. In the bench experiments of the device, the maximum positioning error was 1.25 mm, the average relative error was only 0.94% and the grain loss rate was not more than 2.26%. Its result showed that the system algorithm was reliable, the positioning accuracy was high and the plate taking operation can be completed well. The field verification experiment showed that the forward speed of the unit was in the range of 0.4~0.8 m/s, and the total loss rate of harvest was less than 5%. When the forward speed is 0.6 m/s, the minimum harvest loss rate is 2.32%, which indicated the control system meets the requirements of sunflower harvesting operation.

Flexible design of chiroptical response of planar chiral metamaterials using deep learning

Optical chirality is highly demanded for biochemical sensing, spectral detection, and advanced imaging, however, conventional design schemes for chiral metamaterials require highly computational cost due to the trial-and-error strategy, and it is crucial to accelerate the design process particularly in comparably simple planar chiral metamaterials. Herein, we construct a bidirectional deep learning (BDL) network consists of spectra predicting network (SPN) and design predicting network (DPN) to accelerate the prediction of spectra and inverse design of chiroptical response of planar chiral metamaterials. It is shown that the proposed BDL network can accelerate the design process and exhibit high prediction accuracy. The average process of prediction only takes ∼15 ms, which is 1 in 40000 compared to finite-difference time-domain (FDTD). The mean-square error (MSE) loss of forward and inverse prediction reaches 0.0085 after 100 epochs. Over 95.2% of training samples have MSE ≤ 0.0042 and MSE ≤ 0.0044 for SPN and DPN, respectively; indicating that the BDL network is robust in the inverse deign without underfitting or overfitting for both SPN and DPN. Our founding shows great potentials in accelerating the on-demand design of planar chiral metamaterials.

Error Loss Networks.

A novel model called error loss network (ELN) is proposed to build an error loss function for supervised learning. The ELN is similar in structure to a radial basis function (RBF) neural network, but its input is an error sample and output is a loss corresponding to that error sample. That means the nonlinear input-output mapper of the ELN creates an error loss function. The proposed ELN provides a unified model for a large class of error loss functions, which includes some information-theoretic learning (ITL) loss functions as special cases. The activation function, weight parameters, and network size of the ELN can be predetermined or learned from the error samples. On this basis, we propose a new machine learning paradigm where the learning process is divided into two stages: first, learning a loss function using an ELN; second, using the learned loss function to continue to perform the learning. Experimental results are presented to demonstrate the desirable performance of the new method.

Analysis of generalized type-II progressively hybrid Lindley-exponential data and its modeling in physics, engineering, and management

The life test is guaranteed to end at a predetermined time using a novel type-II generalized progressively hybrid censoring method that is investigated when the test participants’ lifetime model has a two-parameter Lindley-exponential (LndE) distribution. The LndE characteristics are estimated using Bayes and maximum likelihood inference techniques when the suggested censored data are present. By employing the normal approximation of each unknown quantity, the estimated confidence intervals are also constructed. Additionally, independent gamma density priors are used to generate the Bayesian estimators under symmetrical (squared error) loss. Since the likelihood function is formulated in a difficult manner, the Bayes estimators and their corresponding greatest posterior density intervals cannot be computed theoretically, but they may be evaluated using Markov-chain Monte Carlo algorithms. The most progressive design is then determined by applying four optimality criteria. Using Monte Carlo comparisons, the efficiency of the proposed estimating processes is evaluated, and some suggestions are made. In the end, the usefulness of the suggested methods that may be applied in practical situations is demonstrated by analyzing three distinct applications: physics, engineering, and management. When the experimenter’s primary concern is the test’s duration, the numerical findings revealed that the sampling strategy is adaptable and incredibly successful in finishing the experiment in a variety of realistic scenarios.

A comparison framework for deep learning RFI detection algorithms

ABSTRACT In the realm of radio astronomy, the detection of radio frequency interference (RFI) stands as a pivotal pursuit. This study performs a novel comparative analysis of deep learning methodologies and introduces a novel transfer learning method, called fine-tuning. We compare various aspects and factors relating to this problem, including supervised fully convolutional network (FCN) architectures used within the literature, loss functions, regularization techniques, and training methodologies, to establish the most effective strategies for RFI detection. Moreover, the relationship between parameters, FLOPS, and inference times are examined. Fine-tuning involves pre-training models with low-quality aoflagger reference outputs, a very popular and accessible RFI flagging software package, and thereafter re-training the models with high-quality reference outputs. We utilize two data sets: real observations from LOFAR and simulated data from HERA. The mean squared error (MSE) loss function emerges as a robust performer if a high recall is desired. In contrast, the binary cross-entropy (BCE) loss function excels in generalization but falls short in classification performance. The Dice loss function emerges as the top performer, maximizing the F1 score and thereby serving as the choice for our further investigations. Notably, we highlight the important role of data quality and model capacity. In particular, we find that low-capacity models exhibit resilience when trained with low-quality flags from aoflagger, showcasing their ability to mitigate overfitting and overflagging tendencies. In contrast, high-capacity models excel when trained with high-quality flags. Fine-tuning proved to be an effective method to unlearn the overflagging tendencies of aoflagger, whilst requiring very little data.

GAMC: An Unsupervised Method for Fake News Detection Using Graph Autoencoder with Masking

With the rise of social media, the spread of fake news has become a significant concern, potentially misleading public perceptions and impacting social stability. Although deep learning methods like CNNs, RNNs, and Transformer-based models like BERT have enhanced fake news detection. However, they primarily focus on content and do not consider social context during news propagation. Graph-based techniques have incorporated the social context but are limited by the need for large labeled datasets. To address these challenges, this paper introduces GAMC, an unsupervised fake news detection technique using the Graph Autoencoder with Masking and Contrastive learning. By leveraging both the context and content of news propagation as self-supervised signals, our method reduces the dependency on labeled datasets. Specifically, GAMC begins by applying data augmentation to the original news propagation graphs. Subsequently, these augmented graphs are encoded using a graph encoder and subsequently reconstructed via a graph decoder. Finally, a composite loss function that encompasses both reconstruction error and contrastive loss is designed. Firstly, it ensures the model can effectively capture the latent features, based on minimizing the discrepancy between reconstructed and original graph representations. Secondly, it aligns the representations of augmented graphs that originate from the same source. Experiments on the real-world dataset validate the effectiveness of our method.

On bivariate compound exponentiated survival function of the beta distribution: estimation and prediction

A bivariate distribution is a probability distribution that describes the joint behavior of two random variables. It provides information about the simultaneous variation of two variables, allowing us to analyze their relationship and dependencies. This article discussed the bivariate compound exponentiated survival function of the beta distribution. the joint cumulative distribution function and the joint probability density function were found in closed forms. Several characteristics of this distribution have been discussed. The maximum likelihood (ML) estimators of the parameters and two sample ML predictions of the future observations are derived. The Bayes estimators (BEs) of the parameters based on the squared error loss function and two sample Bayesian predictions of the future observations are presented. The performance of the proposed bivariate distributions is examined using a simulation study. Finally, two data sets are considered in the framework of the proposed distributions to demonstrate their flexibility for real-life applications.

MVQS: Robust multi-view instance-level cost-sensitive learning method for imbalanced data classification

Multi-view imbalanced learning is to handle the datasets with multi-view representations and imbalanced classes. Existing multi-view imbalanced learning methods can be divided into two main categories: multi-view ensemble learning and multi-view cost-sensitive learning. However, these methods suffer from the following problems: 1) neglecting either consensus or complementary information, 2) complex preprocessing and information fusion in multi-view ensemble learning and manual assignment of misclassification costs in multi-view cost-sensitive learning, and 3) limited ability to handle noisy samples. Therefore, we aim to design a concise and unified framework to grapple with the multi-view representations, imbalanced classes and noisy samples simultaneously. Inspired by the merits of support vector machine (SVM) and quadratic type squared error (QTSE) loss function, we propose a robust multi-view instance-level cost-sensitive SVM with QTSE loss (MVQS) for imbalanced data classification. The consensus regularization term and combination weight strategy are employed to fully exploit multi-view information. The QTSE loss can adaptively impose instance-level penalties to the misclassification of samples, and make MVQS be robust to noisy samples. We solve MVQS with the alternating direction method of multipliers (ADMM) and the gradient descent (GD) algorithm. Comprehensive experiments validate that MVQS is more competitive and robust than other benchmark approaches.

Estimation of Reliability under Conditional Stress - Strength Setup based on Weibull Distribution

The Weibull distribution has been extensively studied and applied across various fields due to its versatility in modeling a wide range of phenomena, especially in reliability engineering, survival analysis, and lifetime modeling. The concept of Rl a,b , which represents a system's reliability in a conditional stress-strength setup, was proposed by Sabre and Khorshidian (2021). In this research, the problem of estimating reliability of the component is considered when strength variable X and stress variable Y follow independent Weibull distributions with common shapes and different scale parameters under conditional stress-strength setup. The maximum likelihood estimator, asymptotic confidence interval, Bootstrap estimators, Boot-p estimators, and Bayes estimator under-squared error loss function with associated highest posterior density interval are constructed for conditional stress-strength reliability. Simulation study is conducted to estimate mean square error (MSE) of estimator of conditional stress-strength reliability. The real data analysis is also carried out.

Noise aware content-noise complementary GAN with local and global discrimination for low-dose CT denoising

In response to rising concerns over radiation exposure in computed tomography (CT) imaging, effective denoising methods for low-dose CT (LDCT) images are crucial. In recent years, the use of deep learning techniques especially generative adversarial networks (GANs) significantly enhanced the efficiency of LDCT denoising methods, surpassing traditional methods. However, GAN-based denoising methods often face challenges in preserving structural consistency and fine details. This study introduces a novel GAN framework with three accretions to enhance the effectiveness of LDCT denoising. Firstly, our generator is designed to leverage a complementary learning scheme between image noise and image content via two distinct paths. One path focuses on exploring the anatomical information of the image, while the second path is dedicated to learning the noise pattern. This complementary learning scheme provides stable noise cancellation while preserving the maximum structural information of the image. Subsequently, we propose a novel noise-conscious mean absolute error loss to address the challenge posed by the non-stationary characteristic of CT noise. In contrast to conventional MAE loss, this loss attentively prioritizes the different parts of the image based on the local distribution of noise in that region. We also incorporate a gradient-domain loss into the loss function, which inspires the generator to preserve precise image details through explicit guidance. Finally, this study adopted a U-Net-based design for the discriminator to better regularize the model by discriminating between the clean image and the denoised image at both global and local levels. The merit of this discriminator is that it can better adapt to the non-stationary environment of GAN training and guide the generator to produce denoised images that are locally and globally consistent. Our thorough experiments using abdominal CT and lung CT datasets demonstrate the superior performance of our approach compared to state-of-the-art methods.

Optimization of tribological characteristics in cryo-treated plastic/graphene oxide modified CFRP via ANN-based predictive modeling for aerospace applications

Carbon fiber reinforced polymers (CFRP) were modified with polycarbonate (PC)/acrylic butadiene styrene (ABS), and silanized graphene oxide deposited (using electrophoretic deposition) carbon fibers. Using Pin-on-Disc (POD) experiment, the wear rate (WR) and coefficient of friction (COF) of modified CFRP samples were determined for room-temperature (RT) and cryo-treated (CT) conditions. For RT samples, the largest reduction in COF compared to NEAT samples was for SGO samples by 25.6%, while the largest reduction in WR was for PC/ABS samples at 42.1%. But for CT samples, the maximum reduction in COF was 32.6% by SGO, and PC/ABS achieved the maximum reduction in WR by 41%. Also, Analysis of Variance (ANOVA) found “sample composition” to be the most critical component for volume loss mean, whereas Taguchi analysis refined the parameters and achieved the desired outcome. To predict wear behaviour, the Levenberg-Marquardt algorithm, artificial neural network, and linear regression were used. Based on mean squared error (MSE) loss, the Matlab-based Neural Network method performed best, with 0.35% MSE for RT samples and 0.88% for CT samples. This work helps to understand and enhance CFRP materials for wear-resistant applications using optimization techniques.

A deep learning-based imputation method for missing gaps in satellite aerosol products by fusing numerical model data

Satellite-based aerosol optical depth (AOD) products are commonly used in various aerosol-related studies, such as aerosol pollution mapping and aerosol-climate interactions. However, these satellite AOD products often suffer from significant missing gaps due to cloud cover and limitations in the retrieval algorithm. To address this issue, some studies take advantage of real-time seamless simulation of numerical models and successfully fill in these gaps by establishing a regression relationship between satellite AOD and numerical model AOD. However, these previous studies usually use satellite AOD retrievals as the regression target, which limits the accuracy of the imputation results by the original accuracy of satellite AOD retrievals and also consumes a considerable amount of time. To overcome these limitations, this study proposes a spatiotemporal imputation model called Bi-ConvRNN, which combines convolutional neural networks (CNN) and bidirectional recurrent neural networks (Bi-RNN). The model takes both satellite AOD retrievals and numerical model AOD data as input and utilizes the weighted mean squared error (MSE) loss function of multiple AOD datasets, e.g., ground-based data, satellite retrievals, and numerical simulation, as the optimization target to improve the imputation accuracy. The proposed model is evaluated using hourly COMS GOCI AOD products. In the independent test set, the AOD results generated by the Bi-ConvRNN model in the region containing GOCI AOD retrievals can break the accuracy of original GOCI AOD products with the accuracy improved from R2 = 0.70 [RMSE = 0.15] to R2 = 0.84 [RMSE = 0.11], and the filling accuracy, e.g. R2 = 0.79, [RMSE = 0.14], in the region without GOCI AOD retrievals are still better than those of the original GOCI AOD retrievals. Additionally, the Bi-ConvRNN model demonstrates satisfactory filling efficiency, requiring only 0.12 s to fill in the missing gaps of hourly GOCI AOD products per day. These results highlight the efficiency and reliability of the proposed model in filling the gaps in satellite AOD products, and the filled AOD results have great potential for further aerosol-related research.

Classical and Bayesian inference of inverted modified Lindley distribution based on progressive type-II censoring for modeling engineering data

This paper investigates statistical inferences for product lifetimes following the inverted modified Lindley distribution, utilizing progressive Type-II censored data. The estimation of model parameters employs the maximum likelihood method, complemented by the construction of approximate confidence intervals. Bayesian estimates are also explored, incorporating squared error and linear exponential loss functions with noninformative priors. To approximate Bayes estimates, the proposal presents Gibbs sampling based on the MCMC algorithm. This results in the generation of the greatest posterior density credible intervals for the parameters. A real data analysis is conducted to validate the accuracy of all the models and methods discussed. Finally, computational studies using Monte Carlo simulations are presented to compare the suggested estimators.

Accounting journal entries as a long‐term multivariate time series: Forecasting wholesale warehouse output

AbstractLess than 2 years ago, many small entrepreneurs in the commodities trading business faced price volatility, which had not been the case for the last few decades. Generally, the income section of the profit and loss statement was not the main problem, especially in building material commodities trading, due to the recent growth in real estate demand. Logistic disorders, raw material shortages, inflation, and interest rate growth caused difficulties in supply management and warehouse balancing, which were reflected in a particular significant expense called the cost of goods sold. The real problem of its forecasting was identified, and data from accounting books likely contain information about previous warehouse dynamics. This paper presents how accounting data are prepared and shaped into time series suitable for machine learning algorithms, the relevant literature that helped in algorithm selection, and the development and description of the forecasting model, as well as its benchmarking with traditional forecasting models. Visualization and mean squared error loss measured on unseen data show that the model has proven more successful than expected. Based on data from four journal accounts spanning over 14 years, the model predicts the debit and credit sides of the wholesale warehouse for 150 working days.