Root Mean Square Error Research Articles

Robust optimal hybrid fuzzy tuned dual-stage fractional-order PI (1 + PDF) control of fixed-wing aerial robot

This article presents a novel hybrid fuzzy-tuned dual-stage fractional-order (Fo) PI (1 + PDF) controller designed for the robust attitude control of fixed-wing unmanned aerial vehicles (UAVs) under varying aerodynamic uncertainties and wind disturbances. By integrating fractional calculus, dual-stage proportional–integral–derivative (PID) control, and fuzzy logic (FL), the proposed controller aims to enhance performance significantly over traditional methods. The hybrid atom search particle swarm optimization technique is employed to optimize the controller parameters, focusing on minimizing the integral of absolute error (IAE) index. Extensive performance evaluations are conducted across three levels of wind disturbances—light, moderate, and severe—comparing the fuzzy-tuned dual-stage Fo PI (1 + PDF) controller against three alternative strategies: the dual-stage Fo PI (1 + PDF) controller without fuzzy tuning, the dual-stage PI (1 + PDF) controller, and the conventional PID controller. Results demonstrate that under severe wind conditions, the proposed fuzzy-tuned dual-stage Fo PI (1 + PDF) controller significantly enhances control over roll and pitch angles, achieving a 44.7% reduction in IAE, 34.7% improvement in root mean square error (RMSE), and a 65.2% decrease in overshoot for roll control, and a 47.7% reduction in IAE, 42.8% improvement in RMSE, and a 77.1% decrease in overshoot for pitch control, thus validating its efficacy for UAV applications in demanding environmental conditions.

Prediction of pile settlement using a hybrid Adaptive-Network-Based Fuzzy Inference System

Precise pile settlement prediction (SP) in rock-socketed foundations is vital for designing robust bridge foundations and other civil engineering structures. In this work, the behaviors of three powerful algorithms are employed, Dynamic Differential Annealed Optimization (DDAO), Runge Kutta Optimization (RKO), and Ant Lion Optimization (ALO) to improve the performance of the Adaptive Neuro-Fuzzy Inference System (ANFIS) model. In the ANFIS model, some critical input parameters include the rock's unconfined compressive strength, pile length, and pile diameter, which predict SP with high accuracy. The primary contribution of this research is its comparative study with optimization techniques applied to the ANFIS model. Results show that the ANFIS model optimized by DDAO algorithm has the lowest Root Mean Square Error (RMSE) and highest coefficient of determination (R 2 ). On the other side, even though the models optimized through RKO and ALO algorithms also have high predictive capabilities, ALO has extra power in generating a set of Pareto-optimal solutions. This will facilitate the engineers in selecting the most appropriate model given specific design requirements and site-specific constraints. The study provides essential development within the geotechnical engineering study by enhancing the SP prediction accuracy. All these can greatly improve the design and reliability of bridge foundations and other major civil engineering structures for performance and long-term stability.

Hybrid Method for Oil Price Prediction Based on Feature Selection and XGBOOST-LSTM

The accurate and stable prediction of crude oil prices holds significant value, providing insightful guidance for investors and decision-makers. The intricate interplay of factors influencing oil prices and the pronounced fluctuations present significant obstacles within the realm of oil price forecasting. This study introduces a novel hybrid model framework, distinct from the conventional methods, that integrates influencing factors for oil price prediction. First, using Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) extract mode components from crude oil prices. Second, using the Adaptive Copula-based Feature Selection (ACBFS), rooted in Copula theory, facilitates the integration of the influencing factors; ACBFS enhances both accuracy and stability in feature selection, thereby amplifying predictive performance and interpretability. Third, low-frequency modes are predicted through an Attention Mechanism-based Long and Short-Term Memory Neural Network (AM-LSTM), optimized using Bayesian Optimization and Hyperband (BOHB). Conversely, high-frequency modes are forecasted using Extreme Gradient Boosting Models (XGboost). Finally, the error correction mechanism further enhances the predictive accuracy. The experimental results show that the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) of the proposed hybrid prediction framework are the lowest compared to the benchmark model, at 0.7333 and 1.1069, respectively, which proves that the designed prediction structure has better efficiency and higher accuracy and stability.

A Study on Correlation of Depth Fixation with Distance Between Dual Purkinje Images and Pupil Size

In recent times, 3D eye tracking methods have been actively studied to utilize gaze information in various applications. As a result, there is growing interest in gaze depth estimation techniques. This study introduces a monocular method for estimating gaze depth using DPI distance and pupil size. We acquired right eye images from eleven subjects and at ten gaze depth levels ranging from 15 cm to 60 cm at intervals of 5 cm. We used a camera equipped with an infrared LED to capture the images. We applied a contour-based algorithm to detect the first Purkinje image and pupil, then used a template matching algorithm for the fourth Purkinje image. Using the detected features, we calculated the pupil size and DPI distance. We trained a multiple linear regression model on data from eight subjects, achieving an R2 value of 0.71 and a root mean squared error (RMSE) of 7.69 cm. This result indicates an approximate 3.15% reduction in error rate compared to the general linear regression model. Based on the results, we derived the following equation: depth fixation = 20.746 × DPI distance + 5.223 × pupil size + 16.495 × (DPI distance × pupil size) + 13.880. Our experiments confirmed that gaze depth can be effectively estimated from monocular images using DPI distance and pupil size.

Soil Moisture Inversion Using Multi-Sensor Remote Sensing Data Based on Feature Selection Method and Adaptive Stacking Algorithm

Soil moisture (SM) profoundly influences crop growth, yield, soil temperature regulation, and ecological balance maintenance and plays a pivotal role in water resources management and regulation. The focal objective of this investigation is to identify feature parameters closely associated with soil moisture through the implementation of feature selection methods on multi-source remote sensing data. Specifically, three feature selection methods, namely SHApley Additive exPlanations (SHAP), information gain (Info-gain), and Info_gain ∩ SHAP were validated in this study. The multi-source remote sensing data collected from Sentinel-1, Landsat-8, and Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTGTM DEM) enabled the derivation of 25 characteristic parameters through sound computational approaches. Subsequently, a stacking algorithm integrating multiple machine-learning (ML) algorithms based on adaptive learning was engineered to accomplish soil moisture prediction. The attained prediction outcomes were then juxtaposed against those of single models, including Random Forest (RF), Adaptive Boosting (AdaBoost), Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), and Categorical Boosting (CatBoost). Notably, the adoption of feature factors selected by the Info_gain algorithm in combination with the adaptive stacking (Ada-Stacking) algorithm yielded the most optimal soil moisture prediction results. Specifically, the Mean Absolute Error (MAE) was determined to be 1.86 Vol. %, the Root Mean Square Error (RMSE) amounted to 2.68 Vol. %, and the R-squared (R2) reached 0.95. The multifactor integrated model that harnessed optical remote sensing data, radar backscatter coefficients, and topographic data exhibited remarkable accuracy in soil surface moisture retrieval, thus providing valuable insights for soil moisture inversion studies in the designated study area. Furthermore, the Ada-Stacking algorithm demonstrated its potency in integrating multiple models, thereby elevating retrieval accuracy and overcoming the limitations inherent in a single ML model.

XGBoost algorithm optimized by simulated annealing genetic algrithm for permeability prediction modeling of carbonate reservoirs

Carbonate reservoir has strong heterogeneity, complex pore structure and poor correlation between porosity and permeability, so the traditional permeability model can not meet the needs of logging interpretation. Taking the carbonate reservoir of Longwangmiao Formation in Moxi block of central Sichuan as an example, this paper proposes to establish a permeability prediction model by using the XGBoost algorithm of simulated annealing genetic algrithm (SA-GA)hybrid optimization. Combined with core data, five permeability sensitive logging curves (CNL, DEN, DT, and GR) are optimized by calculating correlation coefficients, and the permeability prediction model is established based on XGBoost algorithm, and the XGBoost hyperparameters are optimized by using SA-GA. The method is applied to the evaluation of logging permeability in the study area. The results show that the prediction results of SA-GA-XGBoost algorithm are more consistent with the core data. The adjusted is 0.876, and the root mean square error (RMSE) is only 0.142. The prediction accuracy is better than the conventional permeability model and BP neural network model, which meets the industrial requirements of logging evaluation and provides a new idea for oil and gas exploration in carbonate reservoirs.

Design simulation of high-homogeneity portable MRI magnet array using global optimization algorithm and equivalent currents model.

High-field magnetic resonance imaging (MRI) systems offer high sensitivity and resolution but are costly and bulky, limiting their widespread use, particularly in remote areas. Conversely, portable MRI systems have emerged as a complementary technology, promising enhanced accessibility. This study introduces a novel optimization method combining an analytical model with a highly convergent global optimization algorithm to enhance the design of portable MRI permanent magnet arrays. The approach aims to significantly improve the efficiency of the magnet design process, thereby advancing the homogeneity of portable MRI magnet array. The proposed approach begins with the calculation of initial magnetic field distributions using current element principles. This is followed by the development of an advanced analytical model based on matrix algebra. The consistency between the calculated results of the analytical model and the results from finite element method (FEM) simulations is then evaluated to assess the reliability of the magnetic field calculations across various magnet array configurations. The integration of the analytical model with the improved grey wolf optimization (IGWO) algorithm enhances the optimization process, leading to magnet array configurations with improved homogeneity. FEM simulations agree with the analytical model, revealing a computational error with an average root mean square error (RMSE) of 0.4% in the magnetic field map. The calculation speed of analytical model is at least 200 times higher than that using FEM-based software with uncompromised accuracy. The optimization process successfully yields a permanent magnet array with exceptional homogeneity (1080ppm) and strong field strength (79.5 mT) across a 0.2m diameter of spherical volume (DSV). Moreover, this is accomplished while maintaining a lightweight (129kg) and compact design (interior diameter: 0.31m). The IGWO model has been shown to outperform the benchmark genetic algorithm (GA) model, which is currently used for magnet design in MRI. This study introduces a novel optimization method that significantly enhances the design of portable MRI permanent magnet arrays. By integrating an analytical model with the IGWO algorithm, this method enhances the efficiency of magnet design compared to traditional FEM. This method addresses the limitations of traditional magnet optimization techniques, which are often susceptible to local optima. The results indicate that this method can play a crucial role in developing MRI systems with high homogeneity.

Near Infrared Spectroscopy and Machine Learning for Non-Destructive Estimation of Ageing of Komal Chaul

Komal Chaul is a traditional form of a parboiled rice product from rice varieties indigenous to the state of Assam and it constitutes a culturally significant dish served during the auspicious occasions. Ageing impacts its rehydration properties diminishing its value as a no-cooking rice product. To empower the consumers in selecting Komal Chaul of desired rehydration qualities, this study focused on developing a non-destructive tool based on near-infrared spectral data coupled with machine learning (ML) algorithm for distinguishing the aged Komal Chaul. An NIR spectral library of KomalChaul samples was created, covering the spectral range of 740–1050 nm for samples stored for a period up to one year under ambient conditions. The methodology involved spectral preprocessing to enhance data quality, followed by partial least squares (PLS) regression modeling to predict storage time. Statistical metrics, including regression coefficient (R²), relative error percentage (REP), and root mean squared error (RMSE), were used to validate the model. Feature selection based on coefficient weightage was performed to identify key wavelengths contributing to time prediction. Classification models, including LDA, KNN, CART, Naïve Bayes, SVM, and Random Forest, were employed to categorize samples into aging periods of 1, 3, and 6 months. Partial Least Squares (PLS) regression models predicted the ageing time with a validation score R2 of 0.897 and RMSE of 19.41 days. Optimized with wavelength selection, the PLS regression model achieved significant accuracy in estimating the ageing time, with a prediction score R2 of 0.89 and RMSE of 2.01 days. Similarly, using the same approach for cooking quality prediction resulted in satisfactory performance, achieving a validation R² of 0.79. Classification models further enhanced prediction accuracy, with the Random Forest model attaining the highest accuracy of 92% for six-month interval classifications. These results underscore the potential of integrating NIR spectroscopy and machine learning for efficient, non-destructive quality assessment of Komal Chaul, supporting its commercialization as a value-added traditional food product

Modular Square One-Way Function & Square Root Algorithm (Part-2): AI practical approaches for applying the paper results in GPT

This paper is built upon a previous paper entitled as “Modular Square One-Way Function& Square Root Algorithm: Analyzing the algorithm for randomness, regularity schematic (codec system) and vector normalization “. In that paper the modular square one-way function was analyzed yields the quadratic residue pattern numeric analyzation in the result section. Analyzing the integer factorization results leads to un expected schematic regularity regarding the irrational part of the remainder (decimal expansion) of nonperfect square root. Such regularity was surprising as the expected results assumed to be random. Rounding such rational numbers and normalizing it yield to what is innovatively called modular factor symbol similar to Legendre symbol. Such codec pattern has characteristics of Hilbert envelope, skewness around perfect root pattern with Hann window. In GPU, such calculations could be computed fastly using IEEE- 754 [1] standard for rounding irrational part of the nonperfect square (decimal expansion) with floating point as what mentioned in inverse square root [1]. All above, illuminating an idea of the statistical analyzation for the root mean square error (RMSE). RSME is a powerful estimator of the prediction models used in the artificial intelligence AI especially for the reinforcement learning (RL). As a new approach in AI Google DeepMind Researchers looking through regression analysis algorithm tuning and representing the numerical values as discrete tokens for large language model (LLM). Such data set tokenization and tuning algorithm are helpful for the speed and the predictability of the model as it hase been recognized in the Deep Seek.[4] . Up on all above and considering AI as a new evaluation approach, this paper will discuss the implementation aspects of such innovative results in sampling, tokenizing, clustering and compressing the base model of the GPT a long with fine tuning Neural Network (NN) reasoning of the Reinforcement model.

Improving TerraClimate hydroclimatic data accuracy with XGBoost for regions with sparse gauge networks: A case study of the Meknes plateau and the Middle Atlas Causse, Morocco

Abstract Access to reliable hydroclimatic data, including precipitation, temperature, evapotranspiration, and runoff is crucial for effective water resource management, especially in water-stressed regions like Morocco. However, the scarcity of meteorological stations makes data collection difficult. Satellite products offer a promising alternative to these stations for monitoring and forecasting hydroclimatic trends. This study focuses on the Meknes Plateau and the Middle Atlas Causse to assess the reliability of TerraClimate data and explore their optimization using the XGBoost Machine Learning algorithm. Comparative evaluation between measured data and raw TerraClimate data reveals a satisfactory correlation, though data accuracy imperfections persist. Applying the XGBoost algorithm significantly improves the raw TerraClimate data, reducing the average Mean Absolute Error (MAE) across all parameters from 3.08 to 0.29, and the average Root Mean Square Error (RMSE) from 4.84 to 0.46, and increasing the average Nash-Sutcliffe Efficiency (NSE) from 0.82 to 0.99. These improvements validate this approach in enhancing hydroclimatic data quality in the studied region. In conclusion, this study highlights the potential of satellite products, especially TerraClimate, combined with optimization techniques, for example, the XGBoost algorithm, to address hydroclimatic data shortages in water-stressed regions. The results constitute a robust foundation for future initiatives aimed at improving water resource management and resilience to water challenges in Morocco.

Wideband High-Accuracy Microwave Frequency Measurement and Recognition Enabled by Stimulated Brillouin Scattering

We have developed a high-accuracy, wideband microwave frequency measurement (MFM) system based on frequency-to-time mapping (FTTM) enabled by stimulated Brillouin scattering (SBS). This system, now at the prototype stage, utilizes a broadband linear frequency modulated (LFM) pulse signal as the local signal, which is modulated onto the pump lightwave, while the detected signals are modulated onto the probe lightwave. By analyzing the time values at which Brillouin gain or loss occurs in the pump signal, the frequency of the detected signal can be accurately determined. In experiments, the system successfully measured various signal types, including monotone, multitone, LFM, and Costas frequency modulated signals, as well as their combinations, across a frequency range of 1 GHz to 39 GHz. The system achieves a frequency measurement error below 20 MHz, with a root mean square error (RMSE) of 9.74 MHz, and supports an instantaneous bandwidth of up to 12 GHz across C, X, and Ku bands. Furthermore, leveraging Brillouin loss, the system achieves 100% recognition accuracy for nine distinct microwave signal categories, including both individual and composite signals ranging from 0.47 GHz to 19.47 GHz—a first in the field. Notably, this is accomplished at a system sampling rate of only 50 MSa/s, significantly lower than traditional electronic recognition techniques. This breakthrough represents a significant advancement in electronic reconnaissance and warfare, offering a highly effective solution for frequency measurement and signal recognition with reduced computational and hardware requirements.

Improving agricultural commodity allocation and market regulation: a novel hybrid model based on dual decomposition and enhanced BiLSTM for price prediction

As an essential part of daily life, the drastic fluctuations in agricultural commodity prices significantly impact producers’ motivation and consumers’ quality of life, further exacerbating market uncertainty and unsustainability. The ability to scientifically and effectively predict agricultural commodity prices is of great significance for the rational deployment of market mechanisms, the timely adjustment of supply chains, and the promotion of food policy adjustments. This paper proposes a sustainable hybrid model SV-PSO-BiLSTM which integrates Seasonal-Trend decomposition procedure based on Loess (STL), Variational Mode Decomposition (VMD), Particle Swarm Optimization (PSO), and Bidirectional Long Short-Term Memory (BiLSTM) neural networks. This innovative approach first performs seasonal decomposition of the original data using the STL method, then applies the VMD method for double decomposition of the residual components, reconstructs the data based on sample entropy, and finally predicts agricultural commodity market prices using the BiLSTM network model optimized by the PSO algorithm. This paper investigates the market price dynamics of four agricultural commodities (chili, garlic, ginger, and pork) and one agricultural financial derivative (soybean futures). The experimental results indicate that the proposed SV-PSO-BiLSTM hybrid model achieves average values of 0.2241 for root mean square error (RMSE), 0.1665 for mean absolute error (MAE), 0.0207 for mean absolute percentage error (MAPE), and 0.9851 for the coefficient of determination (R2). These results surpass those of other comparative models, demonstrating stronger generalization, reliability, and stability. The research findings can provide effective guidance for the reasonable regulation of agricultural commodity market prices and further promote the healthy and sustainable development of the agricultural commodity industry.

Machine learning vehicle fuel efficiency prediction

To address the challenges associated with fuel consumption in vehicles with low fuel efficiency, several factors must be recognized. Identifying the key factors of fuel efficiency prediction is crucial for making accurate decisions. Therefore, we propose a comprehensive framework that uses machine learning to predict fuel efficiency by integrating various vehicle information. The proposed method comprises a predictive model and analysis framework utilizing key vehicle attributes, such as fuel type, engine displacement, and vehicle grade, to enhance prediction accuracy. We conducted a comparative study using six machine-learning models. To evaluate the machine learning model, MSE (Mean Square Error), RMSE (Root Mean Square Error), MAE (Mean Absolute Error), and R-squared ( Score) were used. We experimented with SHAP(Shapley Additive Explanations), LIME (Local Interpretable Model-agnostic Explanations), and odds ratio analysis to evaluate the impact of various factors on fuel efficiency. We confirmed that the proposed method can predict fuel efficiency. Extra Trees Regressor and Random Forest Regressor demonstrated high prediction accuracy, particularly excelling in capturing nonlinear relationships. We also underscore the importance of identifying markers to support decision-making, offering critical insights into the key factors impacting fuel efficiency predictions.

Parametric Representation of Tropical Cyclone Outer Radical Wind Profile Using Microwave Radiometer Data

The Soil Moisture Active Passive (SMAP) satellite can measure sea surface winds under tropical cyclone (TC) conditions with its L-band microwave radiometer, without being affected by rainfall or signal saturation. Through the statistical analysis of SMAP data, this study aims to develop radial wind profile models for the TC outer area whose distance from TC center is larger than the radius of maximum wind (Rm). A total of 196 TC cases observed by SMAP were collected between 2015 and 2020, and their intensities range from tropical storm to category 5. Based on the wind and radius data, the key model parameters α and β were fitted through the Rankine vortex model and the tangential wind profile (TWP) Gaussian model, respectively. α and β control the rate of change of the tangential wind speed with radius. Subsequently, for the parametric representation of α and β, we extracted some TC wind filed parameters, such as maximum wind speed (Um), Rm, the average wind speed at Rm (Uma), and the average radius of 17 m/s (R17) and examined the relationship between Uma and Um, the relationship between Rm and R17, the relationship between α, Um and Rm, and the relationship between β, Um and Rm. According to the results, the new radial wind profile models were proposed, i.e., SMAP Rankine Model-4 (SRM-4), SMAP Rankine Model-5 (SRM-5), and SMAP Gaussian Model-1 (SGM-1). A significant advantage of these models is that they can simulate average wind distribution through the conversion from Um to Uma. Finally, comparisons were made between the new models and existing SRM-1, SRM-2, and SRM-3, according to the Advanced Microwave Scanning Radiometer 2 (AMSR-2) measurements of 126 TC cases. The results demonstrate that the SRM-4 simulated the radial wind profile best overall, with the lowest root mean-square error (RMSE) of 5.57 m/s, due to replacing the parameter Um with Uma, using Rankine vortex for α parameterization and modeling with adequate data. Moreover, the models outperform in the Atlantic Ocean, with a RMSE of 5.37 m/s. The new models have the potential to make a contribution to the study of ocean surface dynamics and be used for forcing numerical models under TC conditions.

Geographically Aware Air Quality Prediction Through CNN-LSTM-KAN Hybrid Modeling with Climatic and Topographic Differentiation

Air pollution poses a pressing global challenge, particularly in rapidly industrializing nations like China where deteriorating air quality critically endangers public health and sustainable development. To address the heterogeneous patterns of air pollution across diverse geographical and climatic regions, this study proposes a novel CNN-LSTM-KAN hybrid deep learning framework for high-precision Air Quality Index (AQI) time-series prediction. Through systematic analysis of multi-city AQI datasets encompassing five representative Chinese metropolises—strategically selected to cover diverse climate zones (subtropical to temperate), geographical gradients (coastal to inland), and topographical variations (plains to mountains)—we established three principal methodological advancements. First, Shapiro–Wilk normality testing (p < 0.05) revealed non-Gaussian distribution characteristics in the observational data, providing statistical justification for implementing Gaussian filtering-based noise suppression. Second, our multi-regional validation framework extended beyond conventional single-city approaches, demonstrating model generalizability across distinct environmental contexts. Third, we innovatively integrated Kolmogorov–Arnold Networks (KANs) with attention mechanisms to replace traditional fully connected layers, achieving enhanced feature weighting capacity. Comparative experiments demonstrated superior performance with a 23.6–59.6% reduction in Root-Mean-Square Error (RMSE) relative to baseline LSTM models, along with consistent outperformance over CNN-LSTM hybrids. Cross-regional correlation analyses identified PM2.5/PM10 as dominant predictive factors. The developed model exhibited robust generalization capabilities across geographical divisions (R2 = 0.92–0.99), establishing a reliable decision-support platform for regionally adaptive air quality early-warning systems. This methodological framework provides valuable insights for addressing spatial heterogeneity in environmental modeling applications.

Detection of Elusive Rogue Wave with Cross-Track Interferometric Synthetic Aperture Radar Imaging Approach.

Rogue waves are reported to wreck ships and claim lives. The prompt detection of their presence is difficult due to their small footprint and unpredictable emergence. The retrieval of sea surface height via remote sensing techniques provides a viable solution for detecting rogue waves. However, conventional synthetic aperture radar (SAR) techniques are ineffective at retrieving the surface height profile of rogue waves in real time due to nonlinearity between surface height and normalized radar cross-section (NRCS), which is not obvious in the absence of rogue waves. In this work, a cross-track interferometric SAR (XTI-SAR) imaging approach is proposed to detect elusive rogue waves over a wide area, with sea-surface profiles embedding rogue waves simulated using a probability-based model. The performance of the proposed imaging approach is evaluated in terms of errors in the position and height of rogue-wave peaks, the footprint area of rogue waves, and a root-mean-square error (RMSE) of the sea-surface height profile. Different rogue-wave events under different wind speeds are simulated, and the reconstructed height profiles are analyzed to determine the proper ranges of look angle, baseline, and mean-filter size, among other operation variables, in detecting rogue waves. The proposed approach is validated by simulations in detecting a rogue wave at a spatial resolution of 3 m × 3 m and height accuracy of decimeters.

Hot deformation behavior investigation of heat-resistant aluminum matrix composite based on Arrhenius model and machine learning

The heat-resistant aluminum matrix composite (AMC) exhibits excellent thermal performance due to the presence of heat-resistant dispersed nano-phases. Accurately characterizing high-temperature flow stress is essential for comprehending the mechanisms of deformation and improving material workability. To enhance the accuracy of modeling the flow stress for a new heat-resistant AMC during high-temperature processing, a set of isothermal compression tests at elevated temperatures was conducted. This testing was performed on the composite under varying temperature levels (473, 523, 573, 623, and 673 K) and distinct strain rates (0.001, 0.01, and 0.1 s<sup>-1</sup>). To accurately characterize the flow stress of the composite material at high temperatures, three distinct models were devised: (1) an Arrhenius model that includes strain compensation; (2) a back-propagation neural network (BPNN) model; and (3) a BPNN model optimized using a genetic algorithm (GA-BPNN). The strain compensation theory enhances the Arrhenius model’s ability to capture nonlinear characteristics, while the genetic algorithm (GA) optimizes the BPNN model’s parameter settings. The accuracy of each model in describing flow stress was compared to determine their effectiveness. The findings demonstrate that the GA-BPNN model achieved superior fitting accuracy, with a root mean square error (RMSE) of 6.48, accompanied by a coefficient of determination (R<sup>2</sup>) of 0.991 and a mean absolute error (MAE) of 5.4. To evaluate the generalization capabilities of the three models, a new data set was utilized for verification. The generalization capabilities of the three models were verified using a set of new data. The GA-BPNN model demonstrates outstanding generalization capability, achieving the highest prediction accuracy for new datasets, with R<sup>2</sup> = 0.9102, RMSE = 9.09, and MAE = 7.83. Using the GA-BPNN model’s fitting results, a hot processing map was developed, and the optimal processing window (573 to 673 K) was identified. This study serves as a valuable reference for optimizing the processing parameters of heat-resistant AMCs and proposes a novel approach combining strain compensation and machine learning for high-temperature flow stress description. While the current framework demonstrates computational robustness, extending conclusions to composites with significantly different compositions requires further validation.

Predicting beta-glucan content in food using neural networks

Beta-glucan, a bioactive polysaccharide found in foods like oats, barley, yeast, and mushrooms, is widely recognized for its significant health benefits, including immune modulation and cholesterol reduction. Accurate prediction of beta-glucan content in food products is essential for optimizing formulations, improving quality control, and enhancing dietary recommendations. This study presents a machine learning approach utilizing neural networks to predict beta-glucan content based on key physicochemical and structural attributes of food samples. Input features such as food type, molecular weight, solubility, moisture content, processing conditions, and chemical composition were analyzed and used to train a feedforward neural network. The model was optimized through hyperparameter tuning and validated using cross-validation techniques. Performance evaluation metrics, including R2, Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE), demonstrated the model’s ability to achieve high predictive accuracy and generalizability. The proposed neural network-based framework provides a cost-effective, time-saving alternative to traditional laboratory-based beta-glucan measurements. This predictive tool has practical applications in food manufacturing, nutritional research, and quality assurance, enabling the design of functional foods with enhanced beta-glucan content tailored for specific health benefits.

Comparative Analysis of ANN, GEP, and Water Advance Power Function for Predicting Infiltrated Water Volume in Furrow of Permeable Surface

The present investigation utilized artificial neural networks (ANN) and gene expression programming (GEP) in comparison with the two-point method (TPM) to develop a generalized solution for predicting infiltrated water volume (∀Z) across various soil types under furrow conditions. This work assesses infiltration behavior with respect to experimental data from several temporal contexts. Data distribution and model performance are evaluated via descriptive statistics and correlation tests. Artificial intelligence (AI) models (ANN and GEP) trained and evaluated utilizing input variables—inflow rate (Qin); furrow length (L); waterfront advance time at the end of the furrow (TL); infiltration opportunity time (To); and cross-sectional area of the inflow (Ao) are compared with TPM performance. More precisely and consistently than the water advance power function, AI-based algorithms hope to be invading water volume. Statistical analysis shows that ANN and GEP have lower error metrics, increased generalizability, and better representation of complex infiltration dynamics. The determination coefficient (R2) of ANN data produced 98.1% for testing and 97.8% for validation, while TPM showed accuracy reductions of 2.5% and 4.6%, respectively. On the other side, the R2 of GEP produced 95.7% for testing and 96.1% for validation, while TPM showed accuracy reductions of 0.7% and 3%, respectively. During ANN model computation, TPMs root mean square error (RMSE) of 0.0135 m3/m exceeded all mean values. Errors within 10% relative deviation were displayed using the ANN model ∀Z. Particularly, ANN and GEP, the study revealed that AI techniques predict furrow irrigation penetration of water volume better than the water advance power function. These models advance soil and furrow adaptation, extrapolation, and accuracy. Results show that AI-driven modeling may maximize hydrological assessments and irrigation control.

Exploratory study of predictors of decreased ability to perform activities of daily living in people living with hand osteoarthritis

Objective People with hand osteoarthritis (OA) report decreased ability to perform activities of daily living (ADL). However, few are referred to occupational therapy by their general practitioner. This study aimed to identify clinical predictors of decreased ADL ability in people with hand OA as markers of the need for referral to occupational therapy. Method A cross-sectional study was conducted as an independent add-on to a randomized controlled trial of adults with hand OA (the COLOR trial). Measures of self-reported (ADL Interview) and observed (Assessment of Motor and Process Skills) ADL ability were collected. Data representing potential predictors identified by stakeholders were extracted from the COLOR trial: age, sex, symptom duration, hand OA type, grip strength, pain, stiffness, function, illness perception, and health-related quality of life. Correlational analyses and prediction models were used. Results Correlations between ADL ability and potential predictors in the 62 participants were low to negligible (r < 0.5). Based on root mean square error (RMSE) estimates, prediction models for observed ADL motor (RMSE = 0.3) and ADL process (RMSE = 0.2) ability were more accurate than for self-reported ADL ability (RMSE = 0.6). However, these variables only predicted observed ADL motor and ADL process ability with 16% (adjusted Rs = 0.163) and 12% (adjusted Rs = 0.120) accuracy, respectively. Conclusion The findings suggest that variables representing body functions, perceived health, and quality of life do not predict ADL ability among people living with hand OA. An adequately powered study is recommended to explore this topic further.