Low Error Research Articles

Comparative evaluation of efficacy in working length determination: Radiography versus Electronic apex locators using in vivo and ex vivo methods with stereomicroscope validation

Background: Accurate determination of the working length (WL) is crucial for successful endodontic treatment. Various methods, including radiography and electronic apex locators (EALs), are employed for WL measurement, each with its advantages and limitations. Aim: This study aimed to compare the accuracy of WL determination using conventional radiography and the Root ZX Mini EAL against an ex vivo gold standard method. Materials and Methods: Fifty single-rooted teeth scheduled for extraction were included. WLs were determined using radiography (Grossman’s method) and the Root ZX Mini EAL. An ex vivo method served as the gold standard with WL carried out on extracted teeth under a stereomicroscope. Statistical analysis included Paired samples t-test, Chi-square test, and Bland–Altman plots. Results: The mean WL values were comparable among methods, with slight variations in precision. The Root ZX Mini EAL demonstrated significantly lower mean absolute error compared to radiography (P < 0.001). Accuracy within ± 0.5 mm and ± 1 mm tolerance ranges favored the EAL over radiography (P = 0.04 and P = 0.004, respectively). Conclusion: The Root ZX Mini EAL exhibited superior accuracy and lower error rates in WL determination compared to radiography. Integrating EALs alongside radiographic techniques is recommended to optimize WL precision in clinical practice.

Model reduction of dynamical systems with a novel data-driven approach: The RC-HAVOK algorithm.

This paper introduces a novel data-driven approximation method for the Koopman operator, called the RC-HAVOK algorithm. The RC-HAVOK algorithm combines Reservoir Computing (RC) and the Hankel Alternative View of Koopman (HAVOK) to reduce the size of the linear Koopman operator with a lower error rate. The accuracy and feasibility of the RC-HAVOK algorithm are assessed on Lorenz-like systems and dynamical systems with various nonlinearities, including the quadratic and cubic nonlinearities, hyperbolic tangent function, and piece-wise linear function. Implementation results reveal that the proposed model outperforms a range of other data-driven model identification algorithms, particularly when applied to commonly used Lorenz time series data.

Analisis Kinerja Algoritma Machine Learning dalam Prediksi Harga Cryptocurrency

Cryptocurrency has become an increasingly popular digital asset in recent years. However, cryptocurrency prices are highly volatile and difficult to predict due to being influenced by many factors such as market sentiment, regulations, and technological adoption. This study aims to analyze the performance of several popular machine learning algorithms in accurately predicting cryptocurrency prices. We evaluated four algorithms: Linear Regression, Random Forest, Support Vector Machine (SVM), and Long Short-Term Memory (LSTM) using historical price datasets of Bitcoin, Ethereum, and Litecoin. The data were analyzed by preprocessing steps such as normalization and splitting into training and testing sets. Evaluation metrics used were Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and prediction accuracy. The experimental results showed that the LSTM algorithm had the best performance in predicting cryptocurrency prices with the highest accuracy and lowest error, followed by SVM, Random Forest, and Linear Regression. Further analysis revealed that LSTM was able to capture patterns and trends in complex time series data.

Evaluation of the accuracy of remote emission sensing measurements via real-world vehicle dynamic tests

To evaluate the measurement accuracy of horizontal and vertical remote emission sensing (RES) equipment, a real-world dynamic test was carried out in Chengdu by using electric vehicles equipped with various concentrations of standard gases. In addition, a new vertical remote sensing spectral data algorithm based on image processing (ISPA) was developed to improve the measurement capability. The results showed that the ISPA provided a greater percentage of valid data and lower relative errors; thus, our new algorithm could more effectively analyze the spectral data to measure vehicle emission levels. The percentages of valid horizontal and vertical RES data were 71% and 84%, respectively. The mean relative errors of CO2, CO, HC and NO measured by the vertical RES were about 5%, 20%, 20% and 40%, respectively, and those of CO2, CO and NO measured by the horizontal RES were 3%, 13% and 15%, respectively. For the common vehicle emission concentration, the percentage of valid data for the two RES types increased with increasing gas concentration. As the vehicle speed increased, the relative errors of the horizontal RES equipment showed an increasing trend for the same concentration of gas. Furthermore, for the same speed segment, the relative errors of the horizontal RES equipment increased as the simulated emission concentration decreased. The vertical RES equipment did not exhibit a consistent trend in terms of changes. This study provides a data quality reference for further RES applications.

Model performance evaluation of build time using geometric shape complexity and process parameters in material extrusion

Additive Manufacturing (AM) is gaining widespread interest and recognition in both academia and industry due to its numerous advantages over traditional manufacturing processes. The AM process offers two significant benefits: mass customization and the ability to achieve higher shape complexity at no cost. The present study introduces a methodology for quantifying the shape complexity of feature-based parametric models that AM can produce. This method is exclusively a quantitative approach based on the “number of inputs” required to define the feature or integral sketch. Furthermore, an Artificial Neural Network (ANN) based predictive model for estimating the AM build time has been developed and validated. Supervised Machine Learning (ML) is employed to predict printing time by considering significant design and manufacturing parameters that affect the overall process time. A performance evaluation is conducted between the developed ANN model and a statistical Multiple Linear Regression (MLR) model. The validation Computer Aided Design (CAD) samples are chosen to evaluate the ANN and MLR model's performance to predict the build time. It is observed that the ANN model outperforms the MLR model in estimating the build time of complex geometries. Regarding the efficiency of models, the ANN model achieved an R square value of 0.880 and a lower prediction error of 15.543%. However, the MLR model showed some sensitivity to high complexity and large volume parts when predicting printing time, with a R square value of 0.96. In contrast, the ANN model is more precise in forecasting FDM build time for complex geometries and large volume parts.

Upper limb power-assist wearable robot for handling repetitive medium- to low-weight loads in daily logistics tasks

In this study, we developed an upper-limb power-assisted wearable robot designed to reduce the burden of handling repetitive medium- to low-weight loads for daily logistics workers, thereby enhancing their work efficiency and overall safety. This study proposes a practical wearable robot with a well-designed structure for effectively supporting pick-and-place tasks at waist-to-shoulder height by applying a vertical force directly to the wearer’s wrist. The proposed robot features two active joints that are minimal for vertical assistance, resulting in a lightweight and compact structure. It offers six degrees of freedom per arm, including four passive joints, allowing free end-effector movement. Designed to connect only to the wearer’s wrist, the robot’s linkage is positioned along the wearer’s arm, not requiring alignment with the human–robot joint center, making it easy to wear and having a simple structure. This paper presents a method for calculating the joint torque that accounts for the deformation of the robot’s lightweight and slim links. This approach enhances the gravity compensation accuracy, and the proposed method demonstrates a lower RMS error compared to calculations based on the statics of the rigid link model. Experimental results demonstrated that the robot allowed for a wide range of motion and consistently applied an assistive force of 2 kgf per arm, facilitating the handling of objects weighing several kilograms.

Feature decomposition-based gaze estimation with auxiliary head pose regression

Recognition and understanding of facial images or eye images are critical for eye tracking. Recent studies have shown that the simultaneous use of facial and eye images can effectively lower gaze errors. However, these methods typically consider facial and eye images as two unrelated inputs, without taking into account their distinct representational abilities at the feature level. Additionally, implicitly learned head pose from highly coupled facial features would make the trained model less interpretable and prone to the gaze-head overfitting problem. To address these issues, we propose a method that aims to enhance task-relevant features while suppressing other noises by leveraging feature decomposition. We disentangle eye-related features from the facial image via a projection module and further make them distinctive with an attention-based head pose regression task, which could enhance the representation of gaze-related features and make the model less susceptible to task-irrelevant features. After that, the mutually separated eye features and head pose are recombined to achieve more accurate gaze estimation. Experimental results demonstrate that our method achieves state-of-the-art performance, with an estimation error of 3.90° on the MPIIGaze dataset and 5.15° error on the EyeDiap dataset, respectively.

An unsupervised video anomaly detection method via Optical Flow decomposition and Spatio-Temporal feature learning

The purpose of this paper is to present an unsupervised video anomaly detection method using Optical Flow decomposition and Spatio-Temporal feature learning (OFST). This method employs a combination of optical flow reconstruction and video frame prediction to achieve satisfactory results. The proposed OFST framework is composed of two modules: the Multi-Granularity Memory-augmented Autoencoder with Optical Flow Decomposition (MG-MemAE-OFD) and a Two-Stream Network based on Spatio-Temporal feature learning (TSN-ST). The MG-MemAE-OFD module is composed of three functional blocks: optical flow decomposition, autoencoder, and multi-granularity memory networks. The optical flow decomposition block is used to extract the main motion information of objects in optical flow, and the granularity memory network is utilized to memorize normal patterns and improve the quality of the reconstructions. To predict video frames, we introduce a two-stream network based on spatiotemporal feature learning (TSN-ST), which adopts parallel standard Transformer blocks and a temporal block to learn spatiotemporal features from video frames and optical flows. The OFST combines these two modules so that the prediction error of abnormal samples is further increased due to the larger reconstruction error. In contrast, the normal samples obtain a lower reconstruction error and prediction error. Therefore, the anomaly detection capability of the method is greatly enhanced. Our proposed model was evaluated on public datasets. Specifically, in terms of the area under the curve (AUC), our model achieved an accuracy of 85.74% on the Ped1 dataset, 99.62% on the Ped2 dataset, 93.89% on the Avenue dataset, and 76.0% on the ShanghaiTech Dataset. Our experimental results show an average improvement of 1.2% compared to the current state-of-the-art.

Forecasting the market value of power battery industry chain: A novel RRMIDAS-SVR model

The emergence of electric vehicles has contributed to mitigating air pollution and greenhouse effects caused by traditional fuel vehicles. The power battery industry chain, which is a primary component of electric vehicles, requires more attention to monitor its development status. This study proposes a novel method for forecasting the development status of the power battery industry chain by monitoring the market value index of all listed companies in the power battery industry. It proposes a new forecasting model, RRMIDAS-SVR, which outlines reverse-restricted mixed data sampling (RRMIDAS) into support vector regression (SVR) to end the data-driven challenges of mixed-frequency data and nonlinear relationships. We estimate the RRMIDAS-SVR model using a quadratic programming problem and mixed-frequency West Texas Intermediate crude oil futures prices, electric vehicle sales, and the consumer price index as predictors of the market value of all listed companies in the power battery industry chain. The experimental findings reveal that the RRMIDAS-SVR model outperforms the other models, as evidenced by its lower mean absolute error and root-mean-square error. This study contributes to understanding the development status of the power battery industry value chain by proposing and developing a new approach, RRMIDAS-SVR, to monitor the industry's development status that considers a multi-source information set. Moreover, this study provides strategic insights for stakeholders in the power battery industry.

Tumor spheroid elasticity estimation using mechano-microscopy combined with a conditional generative adversarial network

Background and ObjectivesTechniques for imaging the mechanical properties of cells are needed to study how cell mechanics influence cell function and disease progression. Mechano-microscopy (a high-resolution variant of compression optical coherence elastography) generates elasticity images of a sample undergoing compression from the phase difference between optical coherence microscopy (OCM) B-scans. However, the existing mechano-microscopy signal processing chain (referred to as the algebraic method) assumes the sample stress is uniaxial and axially uniform, such that violation of these assumptions reduces the accuracy and precision of elasticity images. Furthermore, it does not account for prior information regarding the sample geometry or mechanical property distribution. In this study, we investigate the feasibility of training a conditional generative adversarial network (cGAN) to generate elasticity images from phase difference images of samples containing a cell spheroid embedded in a hydrogel. MethodsTo construct the cGAN training and simulated test sets, we generated 30,000 artificial elasticity images using a parametric model and computed the corresponding phase difference images using finite element analysis to simulate compression applied to the artificial samples. We also imaged real MCF7 breast tumor spheroids embedded in hydrogel using mechano-microscopy to construct the experimental test set and evaluated the cGAN using the algebraic elasticity images and co-registered optical coherence microscopy (OCM) and confocal fluorescence microscopy (CFM) images. ResultsComparison with the simulated test set ground truth elasticity images shows the cGAN produces a lower root mean square error (median: 3.47 kPa, 95% confidence interval (CI) [3.41, 3.52]) than the algebraic method (median: 4.91 kPa, 95% CI [4.85, 4.97]). For the experimental test set, the cGAN elasticity images contain features resembling stiff nuclei at locations corresponding to nuclei seen in the algebraic elasticity, OCM, and CFM images. Furthermore, the cGAN elasticity images are higher resolution and more robust to noise than the algebraic elasticity images. ConclusionsThe cGAN elasticity images exhibit better accuracy, spatial resolution, sensitivity, and robustness to noise than the algebraic elasticity images for both simulated and real experimental data.

Fusing multi-scale functional connectivity patterns via Multi-Branch Vision Transformer (MB-ViT) for macaque brain age prediction

Brain age (BA) is defined as a measure of brain maturity and could help characterize both the typical brain development and neuropsychiatric disorders in mammals. Various biological phenotypes have been successfully applied to predict BA of human using chronological age (CA) as label. However, whether the BA of macaque, one of the most important animal models, can also be reliably predicted is largely unknown. To address this question, we propose a novel deep learning model called Multi-Branch Vision Transformer (MB-ViT) to fuse multi-scale (i.e., from coarse-grained to fine-grained) brain functional connectivity (FC) patterns derived from resting state functional magnetic resonance imaging (rs-fMRI) data to predict BA of macaques. The discriminative functional connections and the related brain regions contributing to the prediction are further identified based on Gradient-weighted Class Activation Mapping (Grad-CAM) method. Our proposed model successfully predicts BA of 450 normal rhesus macaques from the publicly available PRIMatE Data Exchange (PRIME-DE) dataset with lower mean absolute error (MAE) and mean square error (MSE) as well as higher Pearson's correlation coefficient (PCC) and coefficient of determination (R2) compared to other baseline models. The correlation between the predicted BA and CA reaches as high as 0.82 of our proposed method. Furthermore, our analysis reveals that the functional connections predominantly contributing to the prediction results are situated in the primary motor cortex (M1), visual cortex, area v23 in the posterior cingulate cortex, and dysgranular temporal pole. In summary, our proposed deep learning model provides an effective tool to accurately predict BA of primates (macaque in this study), and lays a solid foundation for future studies of age-related brain diseases in those animal models.

An adaptive detection framework based on artificial immune for IoT intrusion detection system

Given the continual evolution of new network attack methodologies, defenders face the imperative of constantly upgrading security defenses. Current security technologies, albeit effective against known threats, often fall short in handling the intricacies of diverse and novel attacks. Artificial immunity-based network anomaly detection offers a promising avenue by dynamically adapting to evolving threats. However, prevailing algorithms in this domain suffer from low detection rates, limited adaptability, and extended detector generation times. This study aims to tackle these challenges by introducing a high-efficiency network anomaly detection framework, emphasizing both high-dimensional feature selection and adaptive detector generation. Our approach begins with an enhanced dual-module hybrid high-dimensional feature selection method, leveraging evolutionary principles. Furthermore, we introduce a self-sample clustering algorithm based on fuzzy clustering during the tolerance stage, enhancing detector tolerance efficiency. Additionally, an adaptive detector generation scheme is devised. It divides the non-boundary sub-population based on non-self differences and evolution, while employing the red fox optimization algorithm in the boundary region. This adaptive approach dynamically adjusts detector positions and radii to derive optimal detectors. Through comprehensive validation using well-established IoT and network anomaly datasets, our proposed artificial immunity-based IoT intrusion detection framework exhibits superior performance. It achieves higher classification accuracy and lower error rates compared to current state-of-the-art machine learning and artificial immunity algorithms.

Remaining useful life prediction of Lithium-ion batteries using spatio-temporal multimodal attention networks

Lithium-ion batteries are widely used in various applications, including electric vehicles and renewable energy storage. The prediction of the remaining useful life (RUL) of batteries is crucial for ensuring reliable and efficient operation, as well as reducing maintenance costs. However, determining the life cycle of batteries in real-world scenarios is challenging, and existing methods have limitations in predicting the number of cycles iteratively. In addition, existing works often oversimplify the datasets, neglecting important features of the batteries such as temperature, internal resistance, and material type. To address these limitations, this paper proposes a two-stage RUL prediction scheme for Lithium-ion batteries using a spatio-temporal multimodal attention network (ST-MAN). The proposed ST-MAN is to capture the complex spatio-temporal dependencies in the battery data, including the features that are often neglected in existing works. Despite operating without prior knowledge of end-of-life (EOL) events, our method consistently achieves lower error rates, boasting mean absolute error (MAE) and mean square error (MSE) of 0.0275 and 0.0014, respectively, compared to existing convolutional neural networks (CNN) and long short-term memory (LSTM)-based methods. The proposed method has the potential to improve the reliability and efficiency of battery operations and is applicable in various industries.

Testing the power of Google DeepMind: Gemini versus ChatGPT 4 facing a European ophthalmology examination

PurposeThe aim of this study was to compare the performances of Google Gemini and ChatGPT-4, facing a triple simulation of the European Board of Ophthalmologists (EBO) multiple choices exam. DesignObservational study. MethodsThe EBO multiple choice examination consists of 52 questions followed by 5 statements each, for a total of 260 answers. Statements may be answered with “True”, “False” or “Don't Know”: a correct answer is awarded 1 point; an incorrect is penalized 0.5 points; “don't know” scores 0 points. At least 60 % correct answers are needed to pass the exam. After explaining the rules to the chatbots, he entire question with the 5 statements was input. The rate of correct answers and the final score were collected. The exam simulation was repeated 3 times with randomly generated questions. ResultsGoogle Gemini and ChatGPT-4 succeed in EBO exam simulations in all 3 cases, with an average 85.3 ± 3.1 % and 83.3 ± 2.4 % of correct answers. Gemini had a lower error rate compared to ChatGPT (6.7 ± 1.5 % vs. 13.0 ± 2.6 %, p = 0.03), but answered “Don't know” more frequently (8.0 ± 2.7 % vs. 3.7 ± 1.5 %, p = 0.05). Both chatbots scored at least 70 % of correct answers in each exam subspecialty across the 3 simulations. Converting the percentages into points, Gemini scored 213.5 ± 9.3 points on average, compared to 199.8 ± 7.1 points for ChatGPT (p = 0.21). ConclusionsGoogle Gemini and ChatGPT-4 can both succeed in a complex ophthalmology examination on widespread topics, with higher accuracy compared to their former versions, highlighting their evolving importance in educational and informative setting. PrecisGoogle Gemini and ChatGPT-4 were both able to succeed in 3 consecutive exam simulations of the European Board of Ophthalmologists with an average of 85 % and 83 % correct answers, respectively. Google Gemini showed significantly less errors when compared to ChatGPT.

Bayesian neural network and Bayesian physics-informed neural network via variational inference for seismic petrophysical inversion

Bayesian neural network via variational inference (BNN-VI) aims to estimate probability distribution of the training parameters of the neural network, which are difficult to compute with traditional methods, by proposing a family of densities and finding the candidates that are close to the target. Variational inference is a Bayesian method that provides a faster tool than Markov chain Monte Carlo sampling algorithms. However, to provide an accurate estimation of the most likely models, BNN-VI requires sufficient training data, which is not always available in geophysical inverse problems. We propose a method that combines a Bayesian approach with a physics-informed neural networks algorithms via variational inference, namely Bayesian physics-informed neural networks via variational inference (BPINN-VI) and apply it to a geophysical inverse problem for the estimation of petrophysical properties from seismic data. The method is based on a Bayesian network approach in which the objective function is defined by the Kullback–Leibler (KL) divergence. In addition, the physical relations between data (seismic measurements) and model variables (petrophysical properties) are embedded into neural networks through a physics-dependent loss term. We tested the proposed method on a pre-stack seismic dataset from the North Sea with limited direct measurements of the target petrophysical properties from well logs. The outcome is the most likely model of petrophysical properties. Compared to BNN-VI, the BPINN-VI results show higher correlation and R2 coefficients, and lower mean square error as well as lower uncertainty and higher lateral continuity.

Statistical Postprocessing of Week-1 and Week-2 Precipitation Forecasts Over Taiwan

Abstract The predictability of precipitation is hindered by finer-scale processes not captured explicitly in global numerical models, such as convective interactions, cloud microphysics, and boundary layer dynamics. However, there is growing demand across various sectors for medium- (3–10-day) and extended-range (10–30-day) quantitative precipitation forecasts (QPFs) and probabilistic QPFs (PQPFs). This study uses a novel statistical postprocessing technique, APPM, that combines analog postprocessing (AP) with probability matching (PM) to produce week-1 and week-2 accumulated precipitation forecasts over Taiwan. AP searches for historical predictions that closely resemble the current forecast and create an AP ensemble using the observed high-resolution precipitation patterns corresponding to these forecast analogs. Frequency counting and PM are then separately applied to the AP ensemble to produce calibrated and downscaled PQPFs and bias-reduced QPFs, respectively. Evaluation over a 22-year (1999-2020) period shows that raw ensemble forecasts from the GEFS of NOAA/NWS/Environmental Modeling Center, collected for the Sub-seasonal Experiment, are under-dispersive with a wet bias. In contrast, the AP ensemble spread well represents forecast uncertainty, leading to substantially more reliable and skillful probabilistic forecasts. Furthermore, the AP-based PQPF demonstrates superior discrimination ability and yields notably greater economic benefits for a wider range of users, with the maximum economic value increasing by 30% to 50% for the week-2 forecast. Compared to the raw ensemble mean forecast, the calibrated QPF exhibits lower mean absolute error and explains 3–8 times more variance in observations. Overall, the APPM technique significantly improves week-1 and week-2 QPFs and PQPFs over Taiwan.

Multi-dimensional error figuring model for ion beams in X-ray mirrors

X-ray mirrors are core optical elements in synchrotron light sources, which are typically figured by ion beams. However, traditional ion beam figuring (IBF) methods cannot effectively correct both the height and 2D slope errors. Thus, what we believe to be a novel multi-dimensional error figuring model (MEFM) is proposed to integrate the height and 2D slope errors into a linear system of equations for simultaneous multi-dimensional error convergence. The weight values are determined based on specific manufacturing errors. The simulation results show that MEFM can achieve lower 2D slope and height errors than a single height or 2D slope errors in the figuring method. The power spectral density (PSD) curves before and after simulated figuring demonstrate that the low- and mid-frequency errors are corrected simultaneously. Figuring experiments on a 270 mm × 15 mm X-ray reflector showed that the height and 2D slope errors effectively converged. The results of the PSD curves indicate that MEFM is more effective than the traditional height figuring model in correcting mid-frequency errors, which further proves the validity of the solution model. This study provides an algorithmic reference for the IBF of ultra-high precision X-ray reflectors, thereby promoting the development of physics, medicine, materials, and other fields.

An improved autoencoder for denoising acoustic emission signals in rock fracturing

ABSTRACT Rock fracture acoustic emission (AE) signals are commonly used non-destructive testing data in geological exploration, resource exploitation, and engineering fields. However, these signals are often accompanied by noise interference caused by environmental factors. In this study, we propose an enhanced model for denoising rock fracture AE signals, called simplified fully convolutional denoising autoencoder (SFCDAE). This model is based on the denoising autoencoder principle in the field of deep learning neural networks. The SFCDAE model consists of only seven layers, with minimal preprocessing of data input. By comparing denoising performance evaluation indicators, higher peak signal-to-noise ratio (PSNR) and lower root mean square error (RMSE) were achieved. On average, PSNR increased by 5.575% and RMSE decreased by 22.225%. Using simulated environmental noise to validate the model, it was found that the model has good robustness and can remove artefacts from sudden noise. The practical application value of the LSTM classification model was validated using data containing real experimental noise, resulting in a higher classification accuracy of 80.083%. These results indicate that the proposed model has better denoising performance compared to existing intelligent models and has certain practical value.

Seasonal-scale intercomparison of SMAP and fused SMOS-SMAP soil moisture products

Two L-band passive microwave satellite sensors, onboard the Soil Moisture and Ocean Salinity (SMOS) launched in 2009 and Soil Moisture Active Passive (SMAP) launched in 2015, are specifically designed for surface soil moisture (SM) monitoring. The first global continuous fused L-band satellite SM product based on SMOS and SMAP observations (SMOS-SMAP-INRAE-BORDEAUX, the so-called Fused-IB) was recently released to the public. Currently, the performance of Fused-IB has only been evaluated collectively over the entire data records in the study period, without specific evaluation for individual seasons. To fill this gap, this study intercompared the Fused-IB and the enhanced SMAP-L3 version 6 (SMAP-E) SM products against in situ SM data from the International Soil Moisture Network (ISMN) from 2016 to 2020 regarding the whole period and different seasons. We aim to investigate the performance of these two products at different time scales and to explore the potential eco-hydrological factors (i.e., precipitation and vegetation) driving their seasonal variations. Results show that both SM products are in good agreement with the in situ measurements, demonstrating high median correlation (R) and low ubRMSD (median R = 0.70 and ubRMSD = 0.058 m3/m3 for Fused-IB vs. R = 0.68 and ubRMSD = 0.059 m3/m3 for SMAP-E) during 2016–2020. For most land use and land cover (LULC) types, Fused-IB outperformed SMAP-E with higher accuracy and lower errors, particularly in forests, partly due to the advantage of the robust SMAP-IB (SMAP-INRAE-BORDEAUX) algorithm used to generate Fused-IB in forests, which avoids the pronounced saturation effects of vegetation optical depth caused by relying on optical information. Besides, both products had superior performances across most LULC types in summer (JJA) and autumn (SON), yet increased uncertainties were observed in forests, grasslands, and croplands during spring (MAM) and winter (DJF). These uncertainties could be mainly attributed to the effects of vegetation growth in forests, grasslands and croplands, and the interception of water from rainfall events in grasslands. The results of this study can serve as a reference for algorithm developers to enhance the accuracy of SM and thus promote hydro-meteorological applications that benefit from L-band radiometer soil moisture products.

Prediction and analysis of sea surface temperature based on LSTM-transformer model

This paper introduces a novel method for predicting sea surface temperature (SST) using a hybrid LSTM-Transformer model. The study utilizes the ERA5 dataset for SST prediction at six specific locations near China. The LSTM-Transformer model combines the temporal processing capability of LSTM with the efficient data processing power of Transformer, showing superior performance in reducing Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), as well as improving the R-squared (R²) value, compared to standalone LSTM, Transformer, and traditional Logistic Regression (LR) models. This is particularly evident in spring and autumn, indicating its robustness to seasonal changes. The model's performance varies across different geographic locations, with lower prediction errors in low-latitude and open sea areas, attributed to the less complex environmental dynamics compared to continental shelf areas. Overall, the LSTM-Transformer hybrid model presents a significant advancement in SST prediction, providing important implications for fisheries, meteorology, and climate change research.