Nonlinear Relationship Research Articles

Research on Sales Prediction Based on the LSTM - RVFL - Attention Model

In the era of big data, accurate sales volume and price prediction is crucial for enterprises' market decision-making, inventory management, and pricing strategies. Traditional methods often fail to capture complex temporal patterns and nonlinear relationships in sales data, whereas deep learning offers promising solutions. This study proposes an Attention-based RW-FN-BiLSTM hybrid neural network to improve prediction accuracy. Using a 20152023 Hass Avocado Board dataset, covering multiple regions and key variables like the price and sales volume of conventional and organic avocados, the model integrates BiLSTM for long-term dependencies, RW-FN for feature extraction, and an Attention mechanism for adaptive weight allocation to key features. Experimental results show stable convergence, with predicted values closely following actual trends and an error distribution indicating strong generalization. While some deviations exist, further optimization in network structure and feature selection can enhance performance. This study validates the Attention-based RW-FN-BiLSTM model, providing enterprises with a data-driven approach to market trend analysis, inventory optimization, and pricing decisions, helping them make precise strategic choices in a competitive market environment.

WOB and hook load prediction in drilling based on GA-RF

Abstract Weight on bit (WOB) and hook load are important parameters in the drilling process, and accurate application of WOB and hook load can improve drilling efficiency. The traditional mathematical model needs to consider many relevant factors, the model is difficult to establish, and the prediction accuracy of the model is poor. In order to overcome these problems, this study established a GA-RF prediction model for WOB and hook load based on real-time drilling data of a block in Sichuan Basin. Firstly, the data is filtered and denoised, and the correlation analysis of drilling parameters is carried out. It is found that the linear correlation between the WOB and other drilling parameters was not high, while the linear correlation between the hook load and the standpipe pressure is up to 0.82. Then, a random forest (RF) optimized by genetic algorithm (GA) is established according to the data. The prediction model transmits the number of decision trees and the maximum tree depth of GA to the RF. RF model is used to deal with the nonlinear relationship of drilling parameters to WOB and hook load, and is compared with other models. The research results indicate that the GA-RF algorithm outperforms other algorithms in predicting WOB and hook load, with an R2 as high as 0.976 for predicting WOB and up to 0.998 for predicting hook load. A comparison of various prediction models reveals that the GA-RF model demonstrates high predictive accuracy for both WOB and hook load.

Carbon Assets Management Model Based on Virtual Network Cycle

The improvement of the carbon asset management level of enterprises is a long-term and dynamic process, but lacks a scientific and systematic asset operation management system in business activities. To solve the problem of insufficient data response in business activities, a fast and efficient network architecture needs to be established and network management behavior needs to be improved to meet the increasing demand for carbon trading. A deep neural network least squares method is adopted to process the cyclic data changes in high-dimensional space, solving the problem of low efficiency in reinforcement training caused by the rapid growth of the virtualized network space. This paper addresses the market value attributes of carbon assets in the power system and solves the problem of inaccurate decision-making caused by poor data information collection. It uses a deep learning model to process complex nonlinear relationships and the connection process of carbon asset factor collection, transmission, and feedback, using the Internet of Things (IoT) for communication. It designs a virtualized data collection cycle using network virtualization. It optimizes virtual communication nodes with periodic change and normal data transmission from within the IoT using Software Defined Networking (SDN). After the simulation test, we conclude that the research results have improved the integrity of the management model data. The algorithm performance has been improved by 15% compared with traditional algorithms. Besides, the rapid response mechanism built by SDN network management can enjoy exclusive communication links to improve the reliability of data transmission. It has opened up a networked management innovation model for power system carbon asset management and can provide better technical references for different industries.

Construction of a Dynamic Prediction Model for Student Management Based on Deep Belief Networks

The dynamic nature of academic environments necessitates predictive models capable of addressing temporal, spatial, and categorical complexities inherent in student management. Aligning with the thematic scope of “Frontiers in Computer Science”, this research introduces an innovative framework based on Deep Belief Networks (DBNs) to enhance the accuracy and adaptability of student outcome forecasting. Traditional methods often struggle with integrating heterogeneous data sources, capturing nonlinear relationships, and adapting to fluctuating academic conditions, leading to limited scalability and robustness. To overcome these limitations, we propose the Hierarchical Student Outcome Predictor (HSOP) integrated with the Dynamic Academic Optimization Strategy (DAOS). HSOP leverages temporal–spatial reasoning and hierarchical feature encoding to model complex dependencies in student data. Concurrently, DAOS dynamically adjusts model parameters and regularization strategies using real-time data inputs, ensuring stability and relevance across diverse academic scenarios. Experimental validation demonstrates the superiority of our approach in predicting enrollment trends, academic performance, and resource allocation, significantly outperforming traditional models. This research contributes to advancing AI-based solutions in educational management by addressing scalability, adaptability, and predictive accuracy.

Anemia Increases All-Cause Mortality Risk in Stroke Survivors on Antiplatelet Therapy: A Retrospective Cohort Study.

Approximately 20% of patients with stroke are anemic, and previous studies have identified a U-shaped relationship between hemoglobin levels and all-cause mortality in stroke survivors. However, these studies have not specifically focused on patients with stroke taking antiplatelet agents. This study investigates the impact of anemia and hemoglobin (HGB) on mortality in this population. This study included 356 stroke survivors from the National Health and Nutrition Examination Survey 1999-2018 who were taking antiplatelet agents. It analyzed the impact of HGB levels and anemia on all-cause mortality using Cox regression, examined the nonlinear relationship between HGB and mortality through restricted cubic splines (RCS), and illustrated survival over time using Kaplan-Meier survival curves. RCS analysis revealed no nonlinear relationship between HGB and all-cause mortality (P for overall < 0.01, P for nonlinear = 0.36), with lower HGB levels associated with an increased risk of all-cause mortality. Cox regression analysis showed that HGB was negatively associated with mortality risk across all models (Model 4: hazard ratio = 0.81, 95% confidence intervals 0.73-0.91, P < 0.01). Additionally, anemia significantly increased the risk of mortality in all models (Model 4: hazard ratio = 2.05, 95% confidence intervals 1.43-2.95, P < 0.01). Kaplan-Meier survival curves demonstrated that the survival rate in the anemic group was significantly lower than that of the non-anemic group (P < 0.01). In stroke survivors taking antiplatelet agents, anemia is associated with an increased risk of all-cause mortality, while HGB levels are negatively correlated with mortality risk.

Correlation between pan immune inflammation value and testosterone deficiency risk increase

Testosterone deficiency seriously affects male reproductive function, growth and development, and quality of life. There is a certain association between inflammation and testosterone. PIV, a novel immune-inflammatory biomarker, has emerged. However, little is known about the relationship between PIV and serum testosterone. This study aims to investigate the relationship between PIV and serum testosterone. In this cross-sectional study, we analyzed data from 7389 participants in the National Health and Nutrition Examination Survey (NHANES) from 2011 to 2016. Serum total testosterone levels were measured using precise isotope dilution liquid chromatography and tandem mass spectrometry. PIV was calculated as (neutrophil count × monocyte count × platelet count)/lymphocyte count. Weighted t tests or chi-square tests were utilized to analyze the basic characteristics of the population. Weighted logistic regression analysis, smooth-fit curves, threshold effects, and subgroup analysis were conducted to investigate the correlation between the PIV and testosterone deficiency. Using PIV Quartile 1 as the reference, in the fully adjusted model, the odds ratios (OR) and 95% confidence intervals (CI) for Quartile 2 to Quartile 4 participants were 1.14 (0.93, 1.40), 1.28 (0.99, 1.65), and 1.51 (1.18, 1.95), respectively (P for trend < 0.001), with participants in the highest quartile of PIV having a 51% increased risk of testosterone deficiency compared to those in the lowest quartile of PIV. Smooth-fit curves and threshold effect analysis revealed a nonlinear relationship between PIV and testosterone deficiency, with a turning point at 565.89. The subgroup analysis results showed that, except for obesity, there was no statistically significant difference in the relationship between PIV and testosterone deficiency among different subgroups (P > 0.05). Our study results indicate a positive correlation between PIV and the risk of testosterone deficiency. This suggests that PIV may serve as a potential indicator for testosterone deficiency.

A Five-Dimensional Comprehensive Evaluation of the Yellow River Basin’s Water Environment Using Entropy–Catastrophe Progression Method: Implications for Differentiated Governance Strategies

The systematic evaluation of the water environment in the Yellow River basin is a critical scientific basis for achieving the goals of ecological protection and high-quality development. In this study, a five-dimensional comprehensive evaluation framework (“quality–quantity–space–flow–biota”) consisting of 19 indicators was constructed. The entropy method and the catastrophe progression method were innovatively combined to solve the limitations of traditional evaluation models in characterizing the nonlinear relationships within water environment systems. The results indicated that the Yellow River basin’s overall comprehensive water environment index was 0.032, classified as “good”. However, the spatial differentiation is significant, showing a step-by-step degradation characteristic of “upstream &gt; downstream &gt; midstream”. Gansu Province (0.028), Ningxia Hui Autonomous Region (0.026), Shaanxi (0.024), and Shanxi (0.020) were rated as “poor” and urgently need to be regulated. The core problems are water shortage (Gansu, Ningxia), water quality deterioration (Shaanxi, Shanxi), and fragmentation of aquatic space (Shanxi, Shaanxi). The findings of this study provided a quantitative tool for differentiated governance in the Yellow River basin which could directly support the decision-making needs of “zoning control and precise policy implementation” in the “Outline of the Plan for Ecological Protection and High-quality Development of the Yellow River Basin”.

ViT–KAN Synergistic Fusion: A Novel Framework for Parameter-Efficient Multi-Band PolSAR Land Cover Classification

Deep learning has shown significant potential in multi-band Polarimetric Synthetic Aperture Radar (PolSAR) land cover classification. However, the existing methods face two main challenges: accurately modeling the complex nonlinear relationships between multiple bands and balancing classifier parameter efficiency with classification accuracy. To address these challenges, this paper proposes a novel decision-level multi-band fusion framework that leverages the synergistic optimization of the Vision Transformer (ViT) and Kolmogorov–Arnold Network (KAN). This innovative architecture effectively captures global spatial–spectral correlations through ViT’s cross-band self-attention mechanism and achieves parameter-efficient decision-level probability space mapping using KAN’s spline basis functions. The proposed method significantly enhances the model’s generalization capability across different band combinations. The experimental results on the quad-band (P/L/C/X) Hainan PolSAR dataset, acquired by the Aerial Remote Sensing System of the Chinese Academy of Sciences, show that the proposed framework achieves an overall accuracy of 96.24%, outperforming conventional methods in both accuracy and parameter efficiency. These results demonstrate the practical potential of the proposed method for high-performance and efficient multi-band PolSAR land cover classification.

Predictive role of the dietary inflammatory index on stroke risk among hypertensive patients

This study aims to evaluate the association between the dietary inflammatory index (DII) and stroke risk in hypertensive patients. Data were sourced from the National Health and Nutrition Examination Survey (NHANES) spanning 1999–2020, including 23,712 hypertensive patients. DII scores were calculated based on dietary intake data, and stroke diagnoses were determined through self-reported physician diagnoses. The relationship between DII and stroke risk was assessed using multivariable logistic regression models. Dose–response relationships and subgroup differences were explored through stratified analysis and restricted cubic spline (RCS) methods. Key dietary factors associated with stroke were identified using least absolute shrinkage and selection operator (LASSO) regression and incorporated into a risk prediction nomogram model. The model’s discriminatory ability for stroke was evaluated using receiver operating characteristic (ROC) curves. After adjusting for confounding factors, the highest DII quartile was associated with an adjusted odds ratio (OR) of 1.44 (95% CI 1.19, 1.74) for stroke compared to the lowest quartile, and each unit increase in DII was associated with an OR of 1.08 (95% CI 1.04–1.13) for stroke prevalence. The RCS curve demonstrated a nonlinear relationship between DII and stroke, with a turning point at 0.29. The nomogram model based on key dietary factors identified by LASSO regression had an area under the curve (AUC) of 70.93% (95% CI 69.81%–72.06%). There is a nonlinear relationship between DII and stroke risk in hypertensive patients. Given the inherent limitations of a cross-sectional study design, further research is needed to establish causality.

How Does Green Credit Affect Corporate Green Investment Efficiency? A Test Based on Listed Corporations in China’s Heavy Pollution Industry

Green credit significantly aids green industry development and energy transformation. However, can green credit incentivize heavy polluting corporations to accelerate their green transformation? To assess this question, this research analyzed how green credit affects green investment efficiency of heavy polluting corporations. A fixed-effects model was applied to explore the impact, followed by a threshold effect model to assess whether there is a nonlinear relationship under the effect of other factors. The study shows that green credit can significantly improve the green investment efficiency of heavy polluting corporations. From an internal control perspective, this improvement is significant for corporations that are state-owned or have low executive shareholding. From an external regulation perspective, the improvement is significant for the areas with low financial and environmental regulation. Green credit is influenced by the corporate asset–liability ratio and executives’ green thinking; both have non-linear, single-threshold effects on corporate green investment efficiency.

Use of pulse pressure index for cardiovascular outcomes assessment and development of a coronary heart disease model for the elderly

Background and aimsPulse pressure has been commonly used to assess atherosclerosis and cardiovascular outcomes with the defects of fluctuation. However, the pulse pressure index (PPI), a feasible alternative with the advantage of lower fluctuations, has not been sufficiently researched.MethodsThis study included 10,796 participants over 65 years from the National Health and Nutrition Examination Survey 1999–2018. Cox proportional hazards models, restricted cubic splines and subgroup analysis were used to investigate the association between PPI in the elderly and all-cause and cardiovascular mortality. Subsequently, a prediction model for identifying coronary heart disease (CHD) in elderly individuals was developed using three machine learning algorithms. The impact of each feature on CHD was visualized in the optimal model after comparing the performances of the models.ResultsThis study discovered that the highest levels of PPI were associated with a 28% increased all-cause [HR (95%CI) 1.28 (1.16, 1.42), P < 0.001] and a 36% increased risk of cardiovascular mortality [HR (95%CI) 1.36 (1.08, 1.72), P = 0.008] compared with the first quantile of PPI. The nonlinear relationships between PPI and mortality (all-cause: P for Nonlinear = 0.038; cardiovascular: P for Nonlinear = 0.005) were determined using restricted cubic spline curves. Among the three machine-learning models, random forest model showed the best performance (AUC 0.667 (0.638, 0.696)). In descending order of feature importance, PPI came in second place, with a positive relationship with CHD.ConclusionsThis study indicated a positive correlation between PPI and long-term adverse cardiovascular outcomes among the elderly. Notably, PPI has considerable predictive power for recognizing CHD.Graphical

Machine Learning Models for Chlorophyll-a Forecasting in a Freshwater Lake: Case Study of Lake Taihu

Cyanobacteria harmful blooms (Cyano-HABs) have become a globally critical environmental issue, threatening freshwater ecosystems by degrading water quality and posing risks to human and aquatic life. Chlorophyll-a (Chl-a), a key biomarker of bloom intensity, offers crucial insights into algal bloom dynamics. However, predicting Chl-a concentrations remains challenging due to the complex interactions between various environmental factors. This study utilizes machine learning (ML) models to predict Chl-a concentrations, focusing on Lake Taihu in China, a large eutrophic lake that serves as an example of numerous freshwater lakes suffering from Cyano-HABs. The research leverages nine critical water quality parameters—water temperature, pH, dissolved oxygen, turbidity, electrical conductivity permanganate index, ammonia nitrogen, total phosphorus, and total nitrogen—to develop an ensemble ML model using XGBoost, known for its ability to handle nonlinear relationships and integrate multiple variables. The XGBoost model achieved superior predictive accuracy with an R2 value of 0.78 and RMSE of 8.97 mg/m3 on the test set, outperforming traditional models like linear regression, decision trees, multi-layer perceptrons, support vector regression, and random forests. Feature importance analysis identified electrical conductivity, turbidity, and water temperature as the most significant predictors of Chl-a levels. This study further enhances model interpretability through Pearson correlation analysis, which quantifies the relationships between Chl-a concentrations and other water quality factors. Additionally, we employed principal component analysis (PCA), mutual information, Spearman rank correlation coefficients, and SHAP models to analyze feature importance and model interpretability in ML. The model’s robustness was tested across multiple monitoring sites in Lake Taihu, demonstrating its potential for broader application in other eutrophic lakes facing similar environmental challenges. By providing a reliable tool for forecasting Chl-a concentrations, this research contributes to the development of early warning systems that can help mitigate the impacts of Cyano-HABs, aiding in more effective water resource management.

Deep learning and genomic best linear unbiased prediction integration: An approach to identify potential nonlinear genetic relationships between traits.

Deep learning and genomic best linear unbiased prediction integration: An approach to identify potential nonlinear genetic relationships between traits.

The association between modified cardiometabolic index with non-alcoholic fatty liver disease and liver fibrosis: a cross-sectional study

BackgroundCardiometabolic index (CMI) was proposed ten years ago as an indicator combining obesity and dyslipidemia. This study aimed to investigate the relationships between newly modified CMI (MCMI) with non-alcoholic fatty liver disease (NAFLD) and liver fibrosis.MethodsThis cross-sectional study included participants in the 2017–2018 National Health and Nutrition Examination Survey database (NHANES). Linear regression was used to explore the relationship between MCMI and Baseline characteristics. Logistic regression was conducted to analyze the correlation among MCMI with NAFLD and liver fibrosis. Furthermore, restricted cubic spline (RCS) was performed to estimate nonlinear relationships. Receiver operating characteristic curve (ROC) was used to assess the diagnostic performance of MCMI for NAFLD and liver fibrosis.ResultsA total of 1385 participants were enrolled in the study. After adjusting covariates, participants with high MCMI were related to increased risk of NAFLD (OR = 3.52, 95%CI: 1.44–8.61), compared with those having low MCMI. A linear association was observed between MCMI and NAFLD (p for nonlinear = 0.074), and a J-shaped nonlinear relationship was found between MCMI and liver fibrosis (p for nonlinear = 0.002). The area under the curve (AUC) for MCMI to identify NAFLD was 0.821 (95% CI 0.799–0.843), which was higher than that of CMI (AUC = 0.761, 95%CI: 0.735–0.786), fatty liver index for the U.S. population (USFLI, AUC = 0.799, 95%CI: 0.776–0.822), Triglyceride glucose index (TyG, AUC = 0.738, 95%CI: 0.712–0.765), NAFLD liver fat score (NLFS, AUC = 0.786, 95%CI: 0.761–0.810) and hepatic steatosis index (HSI, AUC = 0.799, 95%CI: 0.775–0.822).ConclusionsThe novel MCMI was positively corelated to the risk of NAFLD. In addition, MCMI was an effective predictor for both NAFLD and liver fibrosis.

MAFLD mediates the association between CHR and gallstones in the U.S. adults: evidence from NHANES 2021–2023

BackgroundGallstones, a global hepatobiliary disorder, are linked to systemic inflammation, lipid disturbances, and metabolic-associated fatty liver disease (MAFLD). This population-based study aims to investigate the association of the novel inflammation-lipid composite biomarker high-sensitivity C-reactive protein-to-HDL cholesterol ratio (CHR) with gallstones and evaluate whether MAFLD mediates this relationship.MethodsThis cross-sectional analysis utilized data from the National Health and Nutrition Examination Survey (NHANES, 2021–2023) to assess the correlation between the CHR and gallstone prevalence through weighted logistic regression. To evaluate potential nonlinear relationships and assess heterogeneity across key demographics, restricted cubic splines (RCS) were employed to model the association, complemented by subgroup analyses stratified by age, sex, and other covariates. A mediation analysis was used for elucidating the mediating effects of MAFLD.ResultsAmong 4,078 participants, 432 (10.60%) had gallstones. After adjusting for confounders, each unit increase in CHR was associated with a 165% increased risk of gallstones (OR: 2.65, 95% CI: 1.43–4.93, P = 0.006). The RCS curve demonstrated a nonlinear association between the CHR and gallstones (Poverall < 0.001, Pnonlinear < 0.001). Mediation analysis indicated that MAFLD explained 27.1% of this association.ConclusionsCHR is positively associated with gallstones, with MAFLD partially mediating this relationship. Managing CHR levels and preventing MAFLD may reduce gallstone incidence.

Pseudo-Dynamic Source Characterization for Geometrically Rough Faults Using Machine Learning

ABSTRACT Accurately predicting intensity and variability of ground motions for future large earthquakes is essential for seismic hazard analysis. Dynamic rupture simulations provide insights into the physics of the time-dependent slip evolution on the fault plane, governed by stress and friction laws, but their computational demands limit their widespread application. In contrast, kinematic simulations offer computational efficiency using predefined spatiotemporal slip patterns that, however, lack physical consistency of dynamic models. Pseudo-dynamic (PD) approaches integrate the physical principles of dynamic rupture into a kinematic framework for efficient ground-motion computation. To capture the complex nonlinear relationships between earthquake source parameters, machine learning (ML) approaches offer a suitable alternative to first-order statistics. In this study, we develop an ML framework using Fourier neural operators to capture the intricate interdependencies between earthquake source parameters extracted from dynamic rupture scenarios. The rupture model generator begins with a stochastic slip distribution and a slip-constrained hypocenter location. Assuming a dynamically consistent Yoffe source time function (STF), the ML model then estimates the kinematic source parameters rupture velocity, peak slip velocity, and rise time. We adjust the fine-scale STF shape to account for small-scale rupture variations for improving high-frequency radiation. In addition, we introduce a method to model fault roughness correlated with stress drop for stochastic slip scenarios. Our PD rupture generator reproduces the mean and standard deviation of ground-motion models for different intensity measures for Mw 6.5 strike-slip scenarios, showing similar ground-motion variability as dynamic models. We prove the reliability of our approach by modeling strong-motion recordings of the 2000 Mw 6.6 Tottori earthquake. We conclude that our method is successful and computationally efficient to model complex interactions between earthquake source parameters for Mw 6.5–7.0 strike-slip events up to a frequency of 5.75 Hz. Our approach improves the accuracy of ground-motion simulations and thus may help advance seismic hazard assessment.

Exploring neurovascular coupling in stroke patients: insights on linear and nonlinear dynamics using transfer entropy.

The study of neurovascular coupling (NVC), the relationship between neuronal activity and cerebral blood flow, is essential for understanding brain physiology in both healthy and pathological states. Current methods to study NVC include neuroimaging techniques with limited temporal resolution and indirect neuronal activity measures. Methods including electroencephalographic (EEG) data are predominantly linear and display limitations that nonlinear methods address. Transfer Entropy (TE) explores linear and nonlinear relationships simultaneously. This study hypothesizes that complex NVC interactions in stroke patients, both linear and&#xD;nonlinear, can be detected using TE. TE between simultaneously recorded EEG and cerebral blood flow velocity (CBFV) signals was computed and analyzed in three settings: ipsilateral (EEG and CBFV from same hemisphere) stroke and nonstroke, and contralateral (EEG from stroke hemisphere, CBFV from nonstroke hemisphere). A surrogate analysis was performed to evaluate the significance of TE values and to identify the nature of the interactions. The results showed that EEG generally influenced CBFV. There were more linear+nonlinear interactions in the ipsilateral nonstroke setting and in the delta band in ipsilateral&#xD;stroke and contralateral settings. Interactions between EEG and CBFV were stronger on the nonstroke side for linear+nonlinear dynamics. The strength and nature of the interactions were weakly correlated with clinical outcomes (e.g., delta band (p<0.05): infarct growth linear = -0.448, linear+nonlinear = -0.339; NHISS linear = -0.473, linear+nonlinear = -0.457). This study exemplifies the benefits of using TE in linear and nonlinear NVC analysis to better understand the implications of these dynamics in stroke severity.

TEMLI: A High-Resolution Air Temperature Estimation Using Machine Learning and Land Surface Data Across Morocco

Abstract Air temperature (Tair) is an important parameter to understand climate dynamics and support decision-making across various sectors. However, accurately estimating Tair at high spatial resolution remains challenging, particularly in data-sparse regions. This study addresses the research gap that arises from the scarcity of observed temperature data by developing a robust machine learning-based estimation model, TEMLI (Temperature Estimation with ML and Land Input). The approach leverages satellite-derived inputs such as Land Surface Temperature (LST), Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), albedo, and wind speed. The objective is to improve the accuracy of Tair estimation by capturing the nonlinear relationships between these variables. Among the models tested within the TEMLI framework, the Multilayer Perceptron (MLP) demonstrated superior performance, achieving a Root Mean Square Error (RMSE) of less than 1.6 °C. Notably, the model excelled in estimating extreme temperatures, especially in high-altitude regions. Trained using Moroccan weather station data, the model’s generalizability across diverse climatic conditions was also validated. This study concludes that integrating satellite-derived data with machine learning offers a scalable and reliable approach for Tair estimation, particularly in regions with limited in situ observations. The findings have significant implications for advancing climate modeling, monitoring extreme events, and informing decision-making in various climate-sensitive sectors. Graphical Abstract This visual summary offers a concise and accessible overview of the study’s main objectives, methodology, and findings. The TEMLI (Temperature Estimation with Machine Learning and Land Inputs) framework aims to improve air temperature (Tair) estimation at high spatial resolution (1 km) using satellite-derived inputs. The left panel introduces the research context, emphasizing the need for accurate temperature datasets in regions with sparse observations. The central panel outlines the methodological flowchart, detailing the various inputs (LST, NDVI, NDWI, albedo, wind speed), the machine learning models tested, and the selection of the optimal model (MLP), and the evaluation metrics and process. The right panel highlights key results, demonstrating TEMLI’s improved performance over ERA5 products with RMSE &lt; 1.6 °C and R² &gt; 0.9, particularly in high-altitude areas. It also summarizes practical applications, such as enhanced hydrological modeling, improved agricultural climate monitoring, and better temperature-based climate adaptation strategies.

Quantitative Assessment of Age-Friendly Design in Mountainous Urban Community Parks Based on Nonlinear Models: An Empirical Study in Chongqing, China

As China undergoes a rapid transition into an aging society, the mobility challenges faced by the older adults in high-density mountainous cities are exacerbated by complex topographical conditions. Existing studies often overlook the nonlinear relationships and the distinct planning requirements of mountainous versus flat urban areas when evaluating the age-friendliness of community parks. This study aims to identify the critical elements of age-friendly planning in mountainous community parks, validate the applicability of nonlinear models, and quantify the dynamic effects of various factors on older adults’ satisfaction, thereby establishing a foundation for targeted optimization strategies. Utilizing principal component analysis (PCA) on 358 valid questionnaires collected from three representative mountainous community parks in Chongqing, six key planning factors were extracted. An ordered logit model was employed for regression and marginal effect analyses. The findings reveal that (1) the ordered logit model outperforms alternative models in terms of goodness-of-fit and computational efficiency, making it particularly suitable for capturing the nonlinear characteristics of satisfaction; (2) accessibility facilities, spatial and social connectivity, and landscape environment emerge as the primary determinants of older adults’ satisfaction, with improvements in accessibility facilities exerting the most significant impact; and (3) the provision of multifunctional activity spaces and increased social opportunities effectively addresses the psychological and physiological needs of the older adults. This study contributes to the optimization of age-friendly design strategies for community parks in high-density mountainous cities, offering valuable insights for promoting healthy aging and fostering inclusive urban environments.

Optimizing the eco-friendly dyeing of wool and nylon fabrics with Prangos ferulacea (L.) Lindl using artificial intelligence

Eco-friendly dyes provide a sustainable alternative by reducing environmental pollution, enhancing biodegradability, and ensuring safer textile processing. This study optimized the dyeing process for wool, a renewable and biodegradable fiber, and nylon, a durable synthetic fiber, using Prangos ferulacea (L.) Lindl., an environmentally friendly natural dye. Response surface methodology (RSM) was employed to design experiments based on dye concentration, time, pH, and temperature, while artificial neural networks were used for modeling, demonstrating superior predictive accuracy over RSM (minimum R2 = 0.96). Sensitivity analysis identified pH as the most influential factor for wool (50% contribution) and dyeing duration for nylon (46% contribution). A genetic algorithm optimized the dyeing conditions, yielding ideal parameters for wool (80.8 wt.% dye concentration, 110.3 min, pH 6.1, and 80.6 °C) and nylon (68.6 wt.% dye concentration, 102.5 min, pH 5.0, and 95.0 °C). Model validation confirmed a complex nonlinear relationship between dyeing parameters and the K/S value, while the incorporation of natural mordants significantly influenced color strength, modifying the absorption spectrum and colorimetric properties. The integration of artificial intelligence techniques provides deeper insights into dyeing mechanisms, advancing optimized, eco-friendly, and efficient textile dyeing methodologies.