Accurate Estimation Model Research Articles

Testing sufficiency for transfer learning

Modern statistical analysis often encounters high dimensional models but with limited sample sizes. This makes it difficult to estimate high-dimensional statistical models based on target data with limited sample size. Then how to borrow information from another large sized source data for more accurate target model estimation becomes an interesting problem. This leads to the useful idea of transfer learning. Various estimation methods in this regard have been developed recently. In this work, we study transfer learning from a different perspective. Specifically, we consider here the problem of testing for transfer learning sufficiency. We denote transfer learning sufficiency to be the null hypothesis. It refers to the situation that, with the help of the source data, the useful information contained in the feature vectors of the target data can be sufficiently extracted for predicting the interested target response. Therefore, the rejection of the null hypothesis implies that information useful for prediction remains in the feature vectors of the target data and thus calls for further exploration. To this end, we develop a novel testing procedure and a centralized and standardized test statistic, whose asymptotic null distribution is analytically derived. Simulation studies are presented to demonstrate the finite sample performance of the proposed method. A deep learning related real data example is presented for illustration purpose.

Underwater Optical Imaging: Methods, Applications and Perspectives

Underwater optical imaging is essential for exploring the underwater environment to provide information for planning and regulating underwater activities in various underwater applications, such as aquaculture farm observation, underwater topographical survey, and underwater infrastructure monitoring. Thus, there is a need to investigate the underwater imaging process and propose clear and long-range underwater optical imaging methods to fulfill the demands of academia and industry. In this manuscript, we classify the eighteen most commonly used underwater optical imaging methods into two groups regarding the imaging principle, (1) hardware and (2) software-based methods, each with an explanation of the theory, features, and applications. Furthermore, we also discuss the current challenges and future directions for improving the performance of current methods, such as improving the accuracy of underwater image formation model estimation, enlarging the underwater image dataset, proposing comprehensive underwater imaging evaluation metrics, estimating underwater depth and integrating different methods (e.g., hardware- and software-based methods for computational imaging) to promote the imaging performance not only in the laboratory but also in practical underwater scenarios.

An Efficient Filtering Approach for Model Estimation in Sparse Regression

ABSTRACTAs technology advances, the scale of data generated is growing exponentially, bringing huge challenges to data storage and computation. To facilitate the computational cost while maintaining model estimation accuracy, subdata selection become important. Conventional methods, such as LASSO and ridge regression, often focus on feature selection. In contrast, methods of subsampling aim at specifying data points to be extracted. However, these subsampling methods often overlook the full consideration of the role of the response variable and its relationship with predictor variables. In this work, we propose a so‐called filtering approach for model estimation (FAME) method to perform subsampling in combination with feature screening. Compared with existing methods, the proposed method can result in the subdata being smaller in size both in terms of the number of features and observations, and also the computational complexity does not increase. The proposed method can be extended to situations when the predictor is binary, the response is binary, or both are binary. The performance of FAME is evaluated in both numerical studies and real data examples.

Accelerated distributed expectation-maximization algorithms for the parameter estimation in multivariate Gaussian mixture models

Rapid development for modeling big data requires effective and efficient methods for estimating the parameters involved. Although several accelerated Expectation-Maximization algorithms have been developed, there still exist two major concerns: reducing computational cost and improving model estimation accuracy. We propose three distributed-like algorithms for multivariate Gaussian mixture models, which can accelerate speed and improve estimation accuracy. The first algorithm is distributed algorithm, which is used to speed up the calculation of classic algorithms and improve its estimation accuracy by averaging the one-step estimators obtained from distributed operators. The second algorithm is distributed online algorithm, which is a distributed stochastic approximation procedure that performs online updates when reading online data. The final algorithm is called distributed monotonically over-relaxed algorithm, which uses an over-relaxation factor and a distributing strategy to improve the estimation accuracy of multivariate Gaussian mixture models. We investigate the stability, sensitivity, convergence, and robustness of these algorithms in a numerical study. We also apply these algorithms to three real data sets for validation.

O18 UNCERTAINTY QUANTIFICATION OF WEARABLE CUFFLESS BLOOD PRESSURE MEASUREMENT

Background and Objective: Due to aleatoric and epistemic uncertainty, most cuffless blood pressure (BP) estimation models struggle to provide reliable and accurate BP estimations. The purpose of this study is to quantify the uncertainty of wearable cuffless BP estimation so as to reduce the impact of uncertainty on the accuracy of model estimation, and in the meanwhile to provide an estimated uncertainty interval (UI) in addition to the point estimation. Methods: We developed a gradient boosting regression tree (GBRT) model to estimate ambulatory BP and the estimation UI with eight noninvasive features extracted from wearable photoplethysmogram (PPG) and electrocardiogram (ECG) signals. We validated the proposed method with the Microsoft Aurora dataset that was originally collected for the Aurora-BP study. We identified 483 subjects (247 males) with wearable watches collecting PPG, ECG, and other signals, while ambulatory BP was monitored hourly using an oscillometric BP device. We trained the model with quantile loss on 60% of subjects (2954 samples), then calibrated the estimated UI with conformal predication with 24% (2148 samples) of subjects and tested the model with 16% (1658 samples) of the subjects. Results: The mean absolute difference (MAD) in systolic BP (SBP) and diastolic BP (DBP) estimated by the GBRT model were 13.96 mmHg and 9.89 mmHg, respectively. Then, with implementation of conformal prediction with error rates of ϵ=0.05, for the test set samples, the percentage of the estimated UI covering the reference BP during the daytime and nighttime phases can reach 91.67% and 95.89%, respectively. Conclusions: The combination of the GBRT model with conformal predication can quantify the uncertainty in cuffless BP estimations. In addition, the estimated uncertainty quantification interval together with the point estimation is potentially more reliable than single-point estimation, which could benefit better diagnosis and treatment of hypertension.

Global soil respiration predictions with associated uncertainties from different spatio-temporal data subsets

Soil respiration (Rs), the second-largest flux in the global carbon cycle, is a crucial but uncertain component. To improve the understanding of global Rs, we constructed single global models, and specific models classified by climate type, land cover type, year of the data record, and elevation range using the random forest algorithm to predict global Rs values and explore the associated uncertainty in the models. The results showed a similar overall predictive performance for the models, with an R-squared value greater than 0.63; however, significant differences were observed compared to the global Rs estimate (23 Pg C). All the models estimated larger values of Rs than the single global model, mainly owing to imbalances in the sample data on which the prediction models were based. One exception to this result is the land cover model, which estimates a smaller global Rs for 2020 (95.1 Pg C). Overall, the single global model estimates were closer to those obtained for temperate zones owing to differences in the training data distribution, which resulted in smaller global estimates than those of other classification-specific models. Prediction models using observations before 2000 tend to underestimate the global Rs. However, the use of classification-specific Rs models proved helpful in addressing the persistent temporal and spatial imbalances in Rs sampling. Expanding the coverage of Rs records both temporally and spatially and updating the global Rs database promptly would improve the estimation accuracy of global Rs prediction models while enhancing the understanding of the overall global carbon budget and the feedback of soil carbon with regard to climate warming.

Toward Large‐Scale Riverine Phosphorus Estimation Using Remote Sensing and Machine Learning

AbstractPhosphorus pollution is a major water quality issue impacting the environment and human health. Traditional methods limit the frequency and extent of total phosphorus (TP) measurements across many rivers. However, remote sensing can accurately estimate riverine TP; nevertheless, no large‐scale assessment of riverine TP using remote sensing exists. Large‐scale models using remote sensing can provide a fast and consistent method for TP measurement, important for data generalization and accessing extensive spatial‐temporal change in TP. Our study uses remote sensing and machine learning to estimate the TP in rivers in the contiguous United States (CONUS). Initially, we developed a national scale matchup data set for Landsat detectable rivers (river width >30 m) using in situ TP and surface reflectance. We used in situ data from the Water Quality Portal (WQP), alongside water surface reflectance data from Landsat 5, 7, and 8 spanning from 1984 to 2021. Then, we used this data set to develop a machine learning (ML) model using different preprocessing methods and algorithms. We found that using high‐level vegetation in the clustering approach and over‐sampling or under‐sampling our training data in the sampling approach improved our model estimation accuracy. We compared XGBLinear, XGBTree, Regularized Random Forest (RRF), and K‐Nearest neighbors ML algorithms and selected XGBLinear as the best model with an R2 of 0.604, RMSE of 0.103 mg/L, mean average error of 0.83, and NSE of 0.602. Finally, we identified human footprint, elevation, river area, and soil erosion as the main attributes influencing the accuracy of estimated TP from the ML model.

Measurement error–tolerant Poisson regression for Valley Fever incidence prediction

The incidence of Valley Fever can be predicted using a Poisson regression model, which links the expected incidence count to influential covariates with model coefficients. In real-world practice, the estimated model coefficients are sensitive to the inevitable measurement errors in the covariates, which are usually manifested as classical errors and/or Berkson errors. While conventional methods have separately addressed each type of error, the joint consideration of both, referred to as mixture measurement error in this article, is currently absent from the existing literature, leaving model coefficient estimators with significant biases. This article presents a measurement-error tolerant Poisson regression to mitigate the impacts of the mixture measurement error in covariates of health-care data analysis. An error-structureadapted quasi-likelihood estimation method is proposed to enhance the accuracy of model estimation. The effectiveness of the proposed method is demonstrated through a numerical case study built upon a Valley Fever investigation, validating its improved estimation accuracy and robustness.

The fusion of vegetation indices increases the accuracy of cotton leaf area prediction.

Rapid and accurate estimation of leaf area index (LAI) is of great significance for the precision agriculture because LAI is an important parameter to evaluate crop canopy structure and growth status. In this study, 20 vegetation indices were constructed by using cotton canopy spectra. Then, cotton LAI estimation models were constructed based on multiple machine learning (ML) methods extreme learning machine (ELM), random forest (RF), back propagation (BP), multivariable linear regression (MLR), support vector machine (SVM)], and the optimal modeling strategy (RF) was selected. Finally, the vegetation indices with a high correlation with LAI were fused to construct the VI-fusion RF model, to explore the potential of multi-vegetation index fusion in the estimation of cotton LAI. The RF model had the highest estimation accuracy among the LAI estimation models, and the estimation accuracy of models constructed by fusing multiple VIs was higher than that of models constructed based on single VIs. Among the multi-VI fusion models, the RF model constructed based on the fusion of seven vegetation indices (MNDSI, SRI, GRVI, REP, CIred-edge, MSR, and NVI) had the highest estimation accuracy, with coefficient of determination (R2), rootmean square error (RMSE), normalized rootmean square error (NRMSE), and mean absolute error (MAE) of 0.90, 0.50, 0.14, and 0.26, respectively. Appropriate fusion of vegetation indices can include more spectral features in modeling and significantly improve the cotton LAI estimation accuracy. This study will provide a technical reference for improving the cotton LAI estimation accuracy, and the proposed method has great potential for crop growth monitoring applications.

Construction of Remote Sensing Quantitative Model for Biomass of Deciduous Broad-Leaved Forest in Mazongling Nature Reserve Based on Machine Learning

As an important forest type, deciduous broad-leaved forest is crucial for estimating forest carbon sequestration capacity and evaluating forest carbon balance. This study focuses on the natural deciduous broad-leaved forest of Mazongling Nature Reserve in Jinzhai County of China. WorldView-2 images were selected as data source. 36 candidate factors including vegetation indices, texture features, and topographic factors were used for modelling. Three machine learning algorithms (i.e., random forest, k-nearest neighbor, and artificial neural network) were used to establish the optimal quantitative retrieval model for natural deciduous broad-leaved biomass. Results showed that the ANN model was the best predictor with R2 = 0.69 and RMSE = 31.53 (Mg·ha−1). Combining the ANN model with the complete spatial coverage of remote sensing data, we developed a distribution map of natural deciduous broad-leaved biomass in the Mazongling forest farm. The estimated average biomass of the study area was 90.34 ± 47.96 Mg·ha−1. In addition, the influence of light saturation on model accuracy is also discussed. This study confirms that remote sensing data in temporal and spatial space can improve the model estimation accuracy.

Improved estimation of canopy water status in cotton using vegetation indices along with textural information from UAV-based multispectral images

Precise and timely estimation of crop canopy water status is of great importance for precision irrigation. The unmanned aerial vehicle (UAV)-based remote sensing technology has become increasingly popular for crop canopy water status estimation. Nevertheless, the capability of vegetation indices (VIs) along with textural information from high-resolution imagery for estimating cotton canopy water status has been rarely explored. The VIs, texture features (TFs), and texture indices (TIs) were obtained from UAV multispectral images of cotton with different irrigation levels and nitrogen rates in the southern Xinjiang of China. The performances of three machine learning models, i.e. support vector machine (SVM), back-propagation neural network (BPNN) and extreme gradient boosting (XGBoost) were evaluated for estimating canopy equivalent water thickness (CEWT) and crop water stress index (CWSI) throughout the cotton growing season. The results showed that combining vegetation indices and textural information significantly improved the estimation accuracy of models compared to vegetation indices or textural information alone. The XGBoost_VIs + TFs model exhibited the best accuracy in estimating CEWT (R2 = 0.75, RMSE = 0.01 cm, RE = 19.46 % at upper half-canopy level, and R2 = 0.65, RMSE = 0.02 cm, RE = 24.59 % at all-canopy level), while the XGBoost_VIs + TFs + TIs model performed best in predicting CWSI among the models (R2 = 0.90, RMSE = 0.05, RE = 5.84 %). Although CEWT estimation was fair to some extent, CWSI estimation was more applicable for diagnosing cotton water stress. The CWSI maps created from the optimal XGBoost_VIs + TFs + TIs model intuitively reflected the cotton canopy water status under various irrigation levels and nitrogen rates, which could help farmers implement timely and precision irrigation in cotton production.

Online battery model parameters identification approach based on bias-compensated forgetting factor recursive least squares

Accuracy of a lithium-ion battery model is pivotal in faithfully representing actual state of battery, thereby influencing safety of entire electric vehicles. Precise estimation of battery model parameters using key measured signals is essential. However, measured signals inevitably carry random noise due to complex real-world operating environments and sensor errors, potentially diminishing model estimation accuracy. Addressing the challenge of accuracy reduction caused by noise, this paper introduces a Bias-Compensated Forgetting Factor Recursive Least Squares (BCFFRLS) method. Initially, a variational error model is crafted to estimate the average weighted variance of random noise. Subsequently, an augmentation matrix is devised to calculate the bias term using augmented and extended parameter vectors, compensating for bias in the parameter estimates. To assess the proposed method's effectiveness in improving parameter identification accuracy, lithium-ion battery experiments were conducted in three test conditions—Urban Dynamometer Driving Schedule (UDDS), Dynamic Stress Test (DST), and Hybrid Pulse Power Characterization (HPPC). The proposed method, alongside two contrasting methods—the offline identification method and Forgetting Factor Recursive Least Squares (FFRLS)—was employed for battery model parameter identification. Comparative analysis reveals substantial improvements, with the mean absolute error reduced by 25%, 28%, and 15%, and the root mean square error reduced by 25.1%, 42.7%, and 15.9% in UDDS, HPPC, and DST operating conditions, respectively, when compared to the FFRLS method.

Tracking the size of the estimation window in time-series data

PurposeThis paper introduces a novel method, Variance Rule-based Window Size Tracking (VR-WT), for deriving a sequence of estimation window sizes. This approach not only identifies structural change points but also ascertains the optimal size of the estimation window. VR-WT is designed to achieve accurate model estimation and is versatile enough to be applied across a range of models in various disciplines.Design/methodology/approachThis paper proposes a new method named Variance Rule-based Window size Tracking (VR-WT), which derives a sequence of estimation window sizes. The concept of VR-WT is inspired by the Potential Scale Reduction Factor (PSRF), a tool used to evaluate the convergence and stationarity of MCMC.FindingsMonte Carlo simulation study demonstrates that VR-WT accurately detects structural change points and select appropriate window sizes. The VR-WT is essential in applications where accurate estimation of model parameters and inference about their value, sign, and significance are critical. The VR-WT has also helped us understand shifts in parameter-based inference, ensuring stability across periods and highlighting how the timing and impact of market shocks vary across fields and datasets.Originality/valueThe first distinction of the VR-WT lies in its purpose and methodological differences. The VR-WT focuses on precise parameter estimation. By dynamically tracking window sizes, VR-WT selects flexible window sizes and enables the visualization of structural changes. The second distinction of VR-WT lies in its broad applicability and versatility. We conducted empirical applications across three fields of study: CAPM; interdependence analysis between global stock markets; and the study of time-dependent energy prices.

Improving Microstructural Estimation in Time-Dependent Diffusion MRI With a Bayesian Method.

Accurately fitting diffusion-time-dependent diffusion MRI (td-dMRI) models poses challenges due to complex and nonlinear formulas, signal noise, and limited clinical data acquisition. Introduce a Bayesian methodology to refine microstructural fitting within the IMPULSED (Imaging Microstructural Parameters Using Limited Spectrally Edited Diffusion) model and optimize the prior distribution within the Bayesian framework. Retrospective. Involving 69 pediatric patients (median age 6 years, interquartile range [IQR] 3-9 years, 61% male) with 41 low-grade and 28 high-grade gliomas, of which 76.8% were identified within the brainstem or cerebellum. 3 T, oscillating gradient spin-echo (OGSE) and pulsed gradient spin-echo (PGSE). The Bayesian method's performance in fitting cell diameter ( ), intracellular volume fraction ( ), and extracellular diffusion coefficient ( ) was compared against the NLLS method, considering simulated and experimental data. The tumor region-of-interest (ROI) were manually delineated on the b0 images. The diagnostic performance in distinguishing high- and low-grade gliomas was assessed, and fitting accuracy was validated against H&E-stained pathology. T-test, receiver operating curve (ROC), area under the curve (AUC) and DeLong's test were conducted. Significance considered at P < 0.05. Bayesian methodology manifested increased accuracy with robust estimates in simulation (RMSE decreased by 29.6%, 40.9%, 13.6%, and STD decreased by 29.2%, 43.5%, and 24.0%, respectively for , , and compared to NLLS), indicating fewer outliers and reduced error. Diagnostic performance for tumor grade was similar in both methods, however, Bayesian method generated smoother microstructural maps (outliers ratio decreased by 45.3% ± 19.4%) and a marginal enhancement in correlation with H&E staining result (r = 0.721 for compared to r = 0.698 using NLLS, P = 0.5764). The proposed Bayesian method substantially enhances the accuracy and robustness of IMPULSED model estimation, suggesting its potential clinical utility in characterizing cellular microstructure. 3 TECHNICAL EFFICACY: Stage 1.

A Survey of Deep Learning Methods for Estimating the Accuracy of Protein Quaternary Structure Models.

The quality prediction of quaternary structure models of a protein complex, in the absence of its true structure, is known as the Estimation of Model Accuracy (EMA). EMA is useful for ranking predicted protein complex structures and using them appropriately in biomedical research, such as protein-protein interaction studies, protein design, and drug discovery. With the advent of more accurate protein complex (multimer) prediction tools, such as AlphaFold2-Multimer and ESMFold, the estimation of the accuracy of protein complex structures has attracted increasing attention. Many deep learning methods have been developed to tackle this problem; however, there is a noticeable absence of a comprehensive overview of these methods to facilitate future development. Addressing this gap, we present a review of deep learning EMA methods for protein complex structures developed in the past several years, analyzing their methodologies, data and feature construction. We also provide a prospective summary of some potential new developments for further improving the accuracy of the EMA methods.

Recent advances and challenges in protein complex model accuracy estimation

Estimation of model accuracy plays a crucial role in protein structure prediction, aiming to evaluate the quality of predicted protein structure models accurately and objectively. This process is not only key to screening candidate models that are close to the real structure, but also provides guidance for further optimization of protein structures. With the significant advancements made by AlphaFold2 in monomer structure, the problem of single-domain protein structure prediction has been widely solved. Correspondingly, the importance of assessing the quality of single-domain protein models decreased, and the research focus has shifted to estimation of model accuracy of protein complexes. In this review, our goal is to provide a comprehensive overview of the reference and statistical metrics, as well as representative methods, and the current challenges within four distinct facets (Topology Global Score, Interface Total Score, Interface Residue-Wise Score, and Tertiary Residue-Wise Score) in the field of complex EMA.

State of Charge and State of Health Coestimation for Lithium-Ion Capacitor Based on Multi-innovation Filters

The lithium-ion capacitor (LIC) is a new type of hybrid energy storage device, which combines the advantages of lithium-ion battery and electric double layer capacitor. To achieve efficient and reliable application of LIC in practical scenarios, accurate model and state estimation method are needed. In this work, the electrical behavior of LIC is studied, which is then described by the Thevenin model. A multi-innovation filter- (MIF-) based coestimation method is proposed, in which the multi-innovation linear Kalman filter (MI-LKF) is used for model parameter identification, the multi-innovation cubature Kalman filter (MI-CKF) is used for state of charge estimation, and the multi-innovation extended Kalman filter (MI-EKF) is used for state of health estimation. Compared to traditional methods, this method can significantly improve estimation accuracy by only expanding the innovation used to update the state from a single moment to multiple moments. The experimental results indicate that the estimation errors of SOC and SOH can be constrained within ±0.5%. In addition, the proposed method has good robustness and can achieve high-precision state estimation even in the face of noise interference, uncertain initial values of the algorithm, and uncertain starting operating points.

Portmanteau Test for ARCH-Type Models by Using High-Frequency Data

The portmanteau test is an effective tool for testing the goodness of fit of models. Motivated by the fact that high-frequency data can improve the estimation accuracy of models, a modified portmanteau test using high-frequency data is proposed for ARCH-type models in this paper. Simulation results show that the empirical size and power of the modified test statistics of the model using high-frequency data are better than those of the daily model. Three stock indices (CSI 300, SSE 50, CSI 500) are taken as an example to illustrate the practical application of the test.

Protein structure accuracy estimation using geometry-complete perceptron networks.

Estimating the accuracy of protein structural models is a critical task in protein bioinformatics. The need for robust methods in the estimation of protein model accuracy (EMA) is prevalent in the field of protein structure prediction, where computationally-predicted structures need to be screened rapidly for the reliability of the positions predicted for each of their amino acid residues and their overall quality. Current methods proposed for EMA are either coupled tightly to existing protein structure prediction methods or evaluate protein structures without sufficiently leveraging the rich, geometric information available in such structures to guide accuracy estimation. In this work, we propose a geometric message passing neural network referred to as the geometry-complete perceptron network for protein structure EMA (GCPNet-EMA), where we demonstrate through rigorous computational benchmarks that GCPNet-EMA's accuracy estimations are 47% faster and more than 10% (6%) more correlated with ground-truth measures of per-residue (per-target) structural accuracy compared to baseline state-of-the-art methods for tertiary (multimer) structure EMA including AlphaFold 2. The source code and data for GCPNet-EMA are available on GitHub, and a public web server implementation is freely available.

Estimating of energy consumption of electric vehicles under different road characteristics: a case study for Nanjing, China

ABSTRACT With the continuous growth of electric vehicle ownership, energy consumption has become an important factor restricting its development. Real-time and accurate energy consumption estimation is essential to alleviate range anxiety and help with energy-efficient route planning. This study aims to explore the factors affecting energy consumption, and estimate the energy consumption of electric vehicles under different road characteristics. Based on the real-world driving data of electric vehicles in Nanjing over a long period, this study analyzed the influencing factors of energy consumption, including trip conditions, weather, driving behavior, and road characteristics. Based on feature selection and correlation analysis, an energy consumption estimation framework for electric vehicles was proposed to estimate energy consumption under different road characteristics using extreme gradient boosting. The results show that energy consumption is significantly affected by trip conditions, weather, driving behavior, and road characteristics. The modeling of different road characteristics helps improve the accuracy of model estimation. The error of the model proposed in this paper is within the reasonable range (−0.25–0.25 kWh/km), which is better than the random forest and k-nearest neighbor models.