Machine Model Research Articles

Development of an artificial intelligence model for wire electrical discharge machining of Inconel 625 in biomedical applications

AbstractSuperalloys, particularly nickel alloys such as Inconel 625, are increasingly used in biomedical engineering for manufacturing critical components such as implants and surgical instruments due to their exceptional mechanical properties and corrosion resistance. However, traditional machining methods often struggle with these materials due to their high strength and thermal conductivity. This study investigates the application of Wire Electrical Discharge Machining (WEDM) as an advanced method for processing Inconel 625 in biomedical contexts. The authors develop an Adaptive Neuro‐Fuzzy Inference System for forecasting WEDM parameters using grey‐based data. The model's variable inputs are analysed through analysis of variance (ANOVA) and Taguchi design, aiming to optimise process performance attributes relevant to biomedical applications. Comparative studies between predicted and experimental data demonstrate a high degree of accuracy, indicating that the proposed model effectively enhances the machining process. The results suggest that this intelligent system supports decision‐making in the production of high‐quality biomedical devices and components.

A Four-Layer Cyber-Physical Security Model for Electric Machine Drives Considering Control Information Flow

A Four-Layer Cyber-Physical Security Model for Electric Machine Drives Considering Control Information Flow

State Machine and Internal Model Control-based Hybrid Controller for Improved Transient Performance of Microgrids

The transient performance of microgrids is measured by their transient stability and transient response indices. The quality of these indices depends upon the design of the microgrid’s closed-loop control scheme, which is a cascaded combination of outer power controller and inner voltage-current (VA) controllers. Generally, the power controller is realized by droop control logic (fixed droop or adaptive droop), and VA controllers by conventional proportional-integral (PI) control. Control schemes with fixed droop coefficient logic and PI-based VA controllers suffer from stability and response issues. In this scheme, when these PI-based VA controllers are replaced with internal model control (IMC) based VA controllers, the response aspect is improved, while the stability aspect shows no improvement. Therefore, when fixed droop logic is replaced with adaptive droop control logic, stability is improved. However, adaptive droop coefficient schemes with PI-based VA controllers provide a limited improvement in stability. Moreover, conventional machine learning-based adaptive droop coefficient logic suffers from severe computational difficulties. It is verified in this work that IMC-based VA controllers which cannot improve stability with fixed droop logic, will contribute to stability enhancement when a suitable adaptive droop control logic is deployed. Therefore, to enhance both stability and response aspects with reduced computational effort, this paper proposes a hybrid control scheme made of a state machine-based droop controller (SMD) supported by IMC-based VA controllers. From the simulation results, while the conventional schemes could not withstand load power factors lesser than 0.707, the proposed hybrid controller scheme maintained stability at a power factor of 0.47 with a total harmonic distortion of less than 5%.

Improving laryngeal cancer detection using chaotic metaheuristics integration with squeeze-and-excitation resnet model.

Laryngeal cancer (LC) represents a substantial world health problem, with diminished survival rates attributed to late-stage diagnoses. Correct treatment for LC is complex, particularly in the final stages. This kind of cancer is a complex malignancy inside the head and neck region of patients. Recently, researchers serving medical consultants to recognize LC efficiently develop different analysis methods and tools. However, these existing tools and techniques have various problems regarding performance constraints, like lesser accuracy in detecting LC at the early stages, additional computational complexity, and colossal time utilization in patient screening. Deep learning (DL) approaches have been established that are effective in the recognition of LC. Therefore, this study develops an efficient LC Detection using the Chaotic Metaheuristics Integration with the DL (LCD-CMDL) technique. The LCD-CMDL technique mainly focuses on detecting and classifying LC utilizing throat region images. In the LCD-CMDL technique, the contrast enhancement process uses the CLAHE approach. For feature extraction, the LCD-CMDL technique applies the Squeeze-and-Excitation ResNet (SE-ResNet) model to learn the complex and intrinsic features from the image preprocessing. Moreover, the hyperparameter tuning of the SE-ResNet approach is performed using a chaotic adaptive sparrow search algorithm (CSSA). Finally, the extreme learning machine (ELM) model was applied to detect and classify the LC. The performance evaluation of the LCD-CMDL approach occurs utilizing a benchmark throat region image database. The experimental values implied the superior performance of the LCD-CMDL approach over recent state-of-the-art approaches.

Deciphering the nonlinear and synergistic role of building energy variables in shaping carbon emissions: A LightGBM- SHAP framework in office buildings

As the global demand for sustainable buildings continues to rise, accurately assessing and managing carbon emissions during the operational phase of buildings has become an urgent challenge. This study employs machine learning and Explainable Artificial Intelligence (XAI) to explore the impact and synergistic effects of energy variables on carbon emissions in office building operations. Initially, energy consumption data from 28 office buildings in China were collected to construct a dataset comprising 24 energy indicators across four dimensions. The Lightweight Gradient Boosting Machine (LightGBM) and Extreme Gradient Boosting (XGBoost) models were used to predict building carbon emission, PV carbon offset, and net carbon emission based on different energy parameters. The SHapley Additive exPlanations (SHAP) method was utilized to interpret the model results from both global and local perspectives. The study results indicate that: (1) The LightGBM model, optimized with GridSearchCV, shows better stability and reliability in predicting carbon emissions. (2) SHAP bar plots和SHAP beeswarm plots reveal that the WWR in the ES dimension and the PVArea in the EP dimension are the most critical variables affecting building carbon emission and PV carbon offset. (3) SHAP dependence plots show that the contributions of energy variables are not independent, with differentiated synergistic effects among variables in emission reduction. (4) SHAP force plots consider interactions among multiple variables, identifying key variable ranges to minimize building carbon emission and maximize PV carbon offset.

Investigating the Impact of Health-Protective Behaviors on Morbidity and non-Morbidity COVID-19 among Workers in Oil Refining Industry Using Machine Learning algorithms

Introduction: The prevalence of COVID-19 has significantly impacted work environments and the workforce. Therefore, identifying the most important preventive and control strategies, as well as assessing their effectiveness, is of paramount importance. Various studies have shown that machine learning algorithms can be used to predict complex and nonlinear issues, including predicting the behavior of various diseases such as COVID-19 and the parameters affecting it, and can be beneficial. The purpose of this study has been to examine the importance of preventive measures and hygiene behaviors in preventing COVID-19 in the oil refining industry using various machine learning models. Material and Methods: For this purpose, demographic information and health behaviors of individuals were collected. Subsequently, a multi-layer perceptron (MLP), radial basis function (RBF), and support vector machine (SVM) models were compared to enhance the analysis of the effects of preventive measures on COVID-19 infection. Finally, the most influential factors affecting the likelihood of COVID-19 infection were determined using sensitivity analysis. Results: The results showed that the accuracies achieved in predicting the impact of preventive measures and health behaviors on COVID-19 in occupational settings were 78.1%, 81.2%, and 78.1% by MLP, RBF, and SVM respectively. The RBF model was identified as the most accurate model for predicting the impact of health behaviors on COVID-19 disease Additionally, the level of social distancing with customers, handwashing frequency and disinfection, the availability of cleansing and disinfecting agents for hands and surfaces in the workplace, and gatherings for eating meals and snacks were identified as the most significant health behaviors influencing the prevalence of COVID-19 in the workplace. Conclusion: Studies of this nature can underscore the importance of attention to preventive measures and health behaviors in unprecedented circumstances. Furthermore, the utilization of artificial intelligence models and tools such as DSS (Decision Support Systems) can serve as powerful tools for optimizing control measures in work environments.

Combining computed tomography features of left atrial epicardial and pericoronary adipose tissue with the triglyceride-glucose index to predict the recurrence of atrial fibrillation after radiofrequency catheter ablation: a machine learning study.

Radiofrequency catheter ablation (RFCA) represents an important treatment option for atrial fibrillation (AF); however, the recurrence rate following surgery is relatively high. This study aimed to predict the recurrence of AF after RFCA using interpretable machine learning models that combined the triglyceride-glucose (TyG) index and the quantification of left atrial epicardial and pericoronary adipose tissue. This retrospective study included 325 patients with AF who underwent their first successful RFCA, among whom 79 had confirmed recurrence. The preoperative clinical data of patients were collected, the TyG index was calculated, and computed tomography (CT) image features were quantitatively measured. Multivariate Cox regression analysis was used to identify the independent risk factors for RFCA recurrence, and adjustments being made for various confounding factors. Post-hoc subgroup analysis was conducted to evaluate the predictive value of the TyG index for recurrence in different patient subgroups. Prediction models based on six machine learning algorithms were constructed. The optimal model's features were evaluated using Shapley additive explanations (SHAP). After adjustment were made for various confounding factors such as comorbidities of AF, Cox regression showed that the volume of left atrial epicardial adipose tissue (LA-EAT), LA-EAT attenuation, left circumflex coronary artery fat attenuation index (LCX-FAI), and the TyG index were independent risk factors for recurrence after RFCA (P<0.001). The support vector machine (SVM) model based on these combined indicators had the best predictive performance, with an area under the curve of 0.793 [95% confidence interval (CI): 0.782-0.805] in the validation set, while its accuracy and positive predictive value were 0.804 and 0.710, respectively. The predictive efficiency of the TyG index appeared to be independent of type 2 diabetes mellitus (T2DM) status (Pinteraction=0.660). The SVM model that integrated the TyG index and quantitative CT imaging variables demonstrated good predictive ability for post-RFCA recurrence in patients with AF. Furthermore, the TyG index appeared capable of predicting recurrence independently of T2DM status.

A large, multi-site lipidomic investigation of parity and aging in dairy cows.

Efforts to optimize the longevity of dairy cows are hindered by the increased risk of adverse health events, culling or dying on farm with increased parity. Lipidomics provides a platform to help identify important biomarkers and biological pathways associated with increased parity and associated aging. A large, multi-site (15 pasture-based, 15 TMR farms) cross-sectional study collected plasma samples from nonlactating, late pregnant, 'dry' cow (696 cows, ~27 d prepartum) and peak-milk cows (796 cows, ~58 DIM) in a disproportionate stratified (parity: 0, 1, 2, > 2 for dry cows; 1, 2, 3, > 3 for peak-milk cows) random sampling frame. A total of 185 lipid species, comprising the lipids classes of phospholipids, sphingomyelins (SM) and triacylglycerols, were quantified in a targeted, liquid chromatography-mass spectrometry (LC-MS) approach. Dry and peak-milk cohorts were analyzed separately throughout. Variation in lipid profiles was mostly attributed to farm of origin (36-41% of variation), with feeding system explaining 13-21% and parity 6-9%, according to ANOVA simultaneous component analysis (ASCA) modeling. Multiple linear regression (MLR) and orthogonal-partial least squares (O-PLS) investigated the association of the lipid profile with age (d), while discriminate analysis compared 1st parity with > 3 parity cows in O-PLS discriminate analysis (O-PLS-DA), random forest (RF) and support vector machine (SVM) models. Rankings of the most important lipid species for each model type were compared. Phospholipids with 40 carbon atoms and 6 double bond equivalents (40:6) were consistently decreased with increasing parity and age across both dry and peak-milk cohorts. These lipids most likely contained stearate (18:0) and docosahexaenoic acid (DHA, C22:6;n-3), an omega-3 fatty acid. Additionally, phospholipids with 40:5, 38:6, lysophosphatidylcholine (17:0), SM(35:1), and SM(35:2) were commonly identified lipids that decreased in concentration with parity and age. Docosahexaenoic acid has been associated with improved cattle health, reproduction, and milk production and quality. This study raises the hypothesis that reduced DHA levels in older cows may be a significant factor increasing susceptibility to adverse health events, reduced reproductive performance, and herd removal. Studies that supplement DHA or its precursors can test this hypothesis and may be important in optimizing longevity of cows.

Self-sensing sliding mode control of workpiece chatter based on accurate prediction of machining vibration

The previous works concerning active chatter suppression in milling systems are generally carried out based on specially designed tool holders or spindles. Due to the continuous removal of materials, however, the workpiece system cannot be treated as quasi-static or rigid. In this paper, a manufacturing system based on the digital twins (DT) model and sliding mode (SM) controller is developed to suppress the milling chatter of thin-walled workpieces. A single degree of freedom (DOF) model of the workpiece is adopted to describe the milling system. The developed SM algorithm exhibits a good performance under the variations of modal parameters, cutting parameters, unmodeled dynamics, etc. To provide accurate control feedback, a DT model of workpiece vibration is established by presenting a self-sensing predictive method. The predictive methods are implemented based on the cooperation of the combination methods of beam functions (CMOBF) and the mode superposition method (MSM). In addition, the predictive accuracy is discussed quantitatively by establishing a dataset of error coefficients. The data of error coefficients are trained by the support vector machine (SVM) model, which is introduced to the predictive methods for further error modification. Finally, a DT framework is introduced briefly to combine the machining system and the control process. Based on the developed DT-driven manufacturing system, the information interaction between the control process and the vibrating system can be realized.

Identifying Hub Genes for Glaucoma based on Bulk RNA Sequencing Data and Multi-machine Learning Models.

The aims of this study were to determine hub genes in glaucoma through multiple machine learning algorithms. Glaucoma has afflicted many patients for many years, with excessive pressure in the eye continuously damaging the nervous system and leading to severe blindness. An effective molecular diagnostic method is currently lacking. The present study attempted to reveal the molecular mechanism and gene regulatory network of hub genes in glaucoma, followed by an attempt to reveal the drug-gene-disease network regulated by hub genes. A microarray sequencing dataset (GSE9944) was obtained through the Gene Expression Omnibus database. The differentially expressed genes in Glaucoma were identified. Based on these genes, we constructed three machine learning models for feature training, Random Forest model (RF), Least absolute shrinkage and selection operator regression model (LASSO), and Support Vector Machines model (SVM). Meanwhile, Weighted Gene Co-Expression Network Analysis (WGCNA) was performed for GSE9944 expression profiles to identify Glaucoma-related genes. The overlapping genes in the four groups were considered as hub genes of Glaucoma. Based on these genes, we also constructed a molecular diagnostic model of Glaucoma. In this study, we also performed molecular docking analysis to explore the gene-drug network targeting hub genes. In addition, we evaluated the immune cell infiltration landscape in Glaucoma samples and normal samples by applying CIBERSORT method. 8 hub genes were determined: ATP6V0D1, PLEC, SLC25A1, HRSP12, PKN1, RHOD, TMEM158 and GSN. The diagnostic model showed excellent diagnostic performance (area under the curve=1). GSN might positively regulate T cell CD4 naïve as well as negatively regulate T cell regulation (Tregs). In addition, we constructed gene-drug networks in an attempt to explore novel therapeutic agents for Glaucoma. Our results systematically determined 8 hub genes and established a molecular diagnostic model that allowed the diagnosis of Glaucoma. Our study provided a basis for future systematic studies of Glaucoma pathogenesis.

Comparative analysis of artificial intelligence models for real-time and future forecasting of environmental conditions: A wood-frame historic building case study

Precise environmental monitoring within historic buildings is crucial for ensuring optimal conditions that guarantee the proper preservation of these structures, thereby maintaining their integrity and cultural legacy. Moreover, outdoor environmental conditions can potentially impact the indoor environment, especially in historic structures with deficient insulation materials and damaged envelopes. As climate change points to more recurrent and severe extreme climatic events in the coming years, a current and future comprehensive evaluation of indoor microclimate in historic structures is essential. This study investigated the utilization of machine learning and deep learning algorithms for real-time and future forecasting of the indoor microclimate, including air temperature, relative humidity, and dew point, in a historic building located in San Antonio, Texas, USA. In situ monitored data from April 2022 to January 2023 were used to train different predictive algorithms. The results indicated that Multi-Layer Perceptrons and Support Vector Machine models yielded the most accurate values in terms of real-time forecasting of the indoor microclimate, while Extreme Gradient Boosting model excelled in convergence time. Additionally, Long Short-Term Memory models were the most accurate in predicting indoor microclimate using future weather data for the current century, specifically for 2050 and 2080. The methodology developed in this study can be applied to different construction types and locations globally. As it enables the prediction of environmental conditions crucial for historic preservation, the results have the potential to assist experts in making informed decisions about conserving historic structures, both now and in the future.

Detection of Fraud and Malware Apps in Google Play

The proliferation of mobile applications on Google Play has led to an increased risk of fraudulent and malicious apps that can compromise user security and privacy. This study presents a comprehensive approach to detecting fraud and malware in Google Play apps using machine learning (ML) and deep learning (DL) techniques. A Flask-based backend is developed to facilitate the upload, processing, and analysis of datasets containing app features and labels. The system utilizes Support Vector Machine (SVM) and Artificial Neural Network (ANN) models to classify apps as benign or malicious. Key stages include data preprocessing, dataset splitting, model training, and performance evaluation. The models' accuracies are compared, and results are visualized to demonstrate their effectiveness. Additionally, an APK upload feature simulates real-time analysis and prediction, enhancing the practical applicability of the solution. The integration of ML and DL methods provides a robust framework for proactively identifying and mitigating threats posed by fraudulent and malware apps in the Google Play ecosystem.

Network pharmacology and transcriptomics reveal androgen receptor as a potential protein target for 6PPD-quinone

N-(1,3-dimethylbutyl)-N′-phenyl-p-phenylenediamine quinone (6PPD-quinone, or 6PPD-Q) has received increasing attention as an emerging hotspot contaminant. The occurrence of 6PPD-Q in dust and fine atmospheric particles indicates substantial human exposure to this toxicant but the hazards of 6PPD-Q to human health is unknown. We used in silico approaches to identify potential human protein targets of 6PPD-Q and conducted preliminary validation through an in vitro cell proliferation assay and an in vivo transcriptomic analysis of prostate tissues from 6PPD-Q-treated mice. Receptor-based reverse screening and network pharmacology identified four hub targets of 6PPD-Q that were closely related to prostate carcinogenesis. Among these four targets, 6PPD-Q exhibited a strong binding tendency to androgen receptor (AR) with a binding free energy of −23.04 kcal/mol. A support vector machine (SVM) model for predicting chemicals with AR agonism or AR-inactivity was established with good prediction performance (mean prediction accuracy: 0.92). SVM prediction and AR-mediated cell-based assays, with a known AR agonist and a proposed AR inactive agent as positive and negative controls, confirmed that 6PPD-Q displayed AR agonism. Upregulation of Ar mRNA expression (FC = 1.29, p = 0.0404) and its related prostate cancer pathway was observed in the prostate of mice exposed to environmentally realistic concentrations of 6PPD-Q, suggesting a potential role in promoting prostate carcinogenesis. These findings provide evidence that 6PPD-Q agonized AR to exert downstream gene transactivation and imply its prostate cancer risks to humans.

Automatic Detection and Classification of Natural Weld Defects Using Alternating Magneto-Optical Imaging and ResNet50.

It is difficult to detect and identify natural defects in welded components. To solve this problem, according to the Faraday magneto-optical (MO) effect, a nondestructive testing system for MO imaging, excited by an alternating magnetic field, is established. For the acquired MO images of crack, pit, lack of penetration, gas pore, and no defect, Gaussian filtering, bilateral filtering, and median filtering are applied for image preprocessing. The effectiveness of these filtering methods is evaluated using metrics such as peak signal-noise ratio (PSNR) and mean squared error. Principal component analysis (PCA) is employed to extract column vector features from the downsampled defect MO images, which then serve as the input layer for the error backpropagation (BP) neural network model and the support vector machine (SVM) model. These two models can be used for the classification of partial defect MO images, but the recognition accuracy for cracks and gas pores is comparatively low. To further enhance the classification accuracy of natural weld defects, a convolutional neural network (CNN) classification model and a ResNet50 classification model for MO images of natural weld defects are established, and the model parameters are evaluated and optimized. The experimental results show that the overall classification accuracy of the ResNet50 model is 99%. Compared with the PCA-SVM model and CNN model, the overall classification accuracy was increased by 7.4% and 1.8%, and the classification accuracy of gas pore increased by 10% and 4%, respectively, indicating that the ResNet50 model can effectively and accurately classify natural weld defects.

On the use of modal works of cutting forces to optimize machining conditions in the presence of vibrations

The use of Virtual Machining models may be a valuable approach in the designing stage of a machining operation as long as the models are sufficiently accurate. When vibration risks are suspected, stability analysis approaches to predict regenerative chatter phenomena are generally used. However, although these approaches, when applicable, allow efficient numerical optimization of the process around an operating point, they often require other strong assumptions such as neglecting transient phenomena or oversimplifying kinematics. On the other hand, time domain approaches with detailed matter removal modelling allow to monitor the continuous evolution of cutting conditions and represent various phenomena that the models can reproduce (regenerative chatter, forced vibrations, non-linear behaviours). The amount of data produced is, however, considerable and often costly to analyse. It may therefore be interesting to deduce, from these data, scalar indicators allowing easier and more relevant analysis of the simulation results.In this work, the modal work of the cutting forces upon the workpiece vibratory displacements is proposed as an indicator to discriminate different tool paths. A one degree of freedom theoretical problem and a face milling operation on extruded aluminum profiles extracted from automotive structural part are used to explain and show the relevance of such indicator.

Landslide susceptibility evaluation and determination of critical influencing factors in eastern Sichuan mountainous area, China

Landslide susceptibility evaluation and determination of critical influencing factors is a prerequisite for preventing hazardous risks, especially in landslide-prone mountainous areas. However, in densely vegetated Southwest mountainous areas, identifying assessment approach of shallow landslides susceptibility and their major inducing factors is still a huge challenge. To address this challenge, we applied five advanced machine learning models (Logistic Regression Model, Generalized Additive Model, Random Forest Model, Support Vector Machine Model, Artificial Neural Network Model) to assess the spatial distribution of shallow landslide susceptibility, considering several relevant factors that affect landslide occurrence. These factors include geological, topographic and vegetation factors, as well as four new vegetation factors: stock volume, stand density, average tree age, and stand types. Furthermore, we employed SHAP algorithm and Structural Equation Models to quantify the relative importance and explanatory power of these factors on shallow landslide susceptibility and to clarify the interaction mechanisms among various factors in Huaying Mountain. The results shown that Random Forest Model proves to be the most accurate (95.1 %) in assessing the spatial distribution of shallow landslides susceptibility, followed by the Artificial Neural Network model (78.6 %), the Support Vector Machine model (69.8 %), the Generalized additive model (68.1 %) and the Logistic Regression model (67.6 %).The area with high susceptible landslide possibility was 25.3 km2 occupying 14.8 % of the study region, it is mainly distributed in the west of Tianchi Lake, southeast of Huaying City and west of the study area, along with Xiangyu Railway. Geographical environment and vegetation features were found to significantly explain 67.4 % and 32.6 % of the total effects in shallow landslides susceptibility, respectively. Specifically, the spatial distribution of shallow landslides susceptibility were primarily influenced by geological engineering rock group, distance to faults、stand types and distance to river. Geographical environment factors could indirectly affect changes in vegetation features, thereby indirectly affecting the spatial distribution of shallow landslides susceptibility. Findings from this research could be helpful for scientific decision-making and technical assistance for early warning, prevention, and control of rainstorm-induced landslides in highly vegetation covered areas.

Fatigue Detection of Air Traffic Controllers Through Their Eye Movements

Eye movement patterns have become an essential element in modern approaches for identifying air traffic controller fatigue. By observing eye movements among various individuals and environments, researchers have discovered correlations with multiple physiological metrics and cognitive processing abilities. This study involved human-in-the-loop simulations to collect eye movement and fatigue data from air traffic controllers and students. The eye movements were classified into three main types: fixation, saccade, and blink. Statistical analyses were performed to determine the most important indicators. Using support vector machine and random forest models for training and prediction, it was found that the fixation characteristic is significantly important for monitoring air traffic controller fatigue. The implementation of this model has the potential to identify forthcoming instances of controller fatigue during their shifts, thereby helping to avert the possibility of unsafe situations.

A deep extreme learning machine approach optimized by sparrow search algorithm for forecasting of traffic flow

Abstract Traffic flow modeling has a pivotal role within Intelligent Transportation Systems (ITSs), holding vital importance in alleviating traffic congestion and decreasing carbon emissions. Due to the presence of variability and nonlinear attributes in traffic flow, developing an effective and resilient model for predicting traffic flow poses a significant challenge. Precisely predicting traffic flow is not merely a feasible issue; it also poses significant difficulties to the researchers involved in this field. This study proposes a hybrid predictive model to forecast traffic flow. The proposed model effectively merges the strengths of the Sparrow Search algorithm (SSA) and Multi-layer Extreme Learning Machine (ML-ELM) model, enhancing prediction accuracy. SSA optimization technique is applied to optimize the initial weights and bias parameters for ML-ELM model. ELM approach is a machine learning approach that employs a single hidden layer to address various tasks. However, in situations where more complex problems are encountered, ML-ELM extends this concept by incorporating multiple hidden layers to enhance its capabilities and address challenges more effectively. Finally, SSA technique is utilized to achieve the optimal tuning of hyperparameters in the context of ML-ELM model to improve the prediction accuracy. Compared to the other selected models, the proposed model outperforms them in terms of performance metrics, including Root Mean Square Errors (RMSE), Mean Absolute Errors (MAE), Mean Absolute Percentage Errors (MAPE) and Correlation Coefficients (r), indicating that it is appropriate for this prediction task.

Modelling the impact of past and future land-use changes on land cover degradation at territorial level in Eastern DR Congo

AbstractLand use and land cover (LULC) changes are recognized as drivers of environmental changes at the local, regional, and global levels. Detecting these changes is essential in developing land management plans and strategies. More particularly, the LULC changes constitute one of the sub-indicators used for setting the baseline for land degradation neutrality (LDN) planning and assessing the progress toward the sustainable development goal 15.3. This study aims to determine the implication of LULC dynamics on land cover degradation in Eastern DR Congo. This region has been subjected to uncontrolled LULC changes over the last three decades but the impact of these changes on land degradation has not been assessed yet at the territorial level. To fill this gap, the territory of Kalehe has been used as a case study to monitor the LULC changes during the 1987–2020 period based on the classification of Landsat images, to forecast the future LULC for the 2030–2070 period through the hybrid Markov-Support Vector Machine modeling, to determine the extent of land cover degradation associated with these changes and to assess the impact of biophysical and socio-economic factors on the occurrence of land cover degradation using the binary logistic regression model. The results reveal that the overall trends of LULC changes during the 1987–2020 period are the increasing of built-up area, shrubland, and cropland at the expense of forestland, wetland, and grassland. This situation is expected to continue in the future and contribute to the degradation of land cover within the study area. Under the current situation of LULC changes, 34.17% of the land has been subjected to potential degradation. Furthermore, under the business-as-usual scenario 28.28%, 27.28%, and 33.65% of the land will be degraded by 2030, 2050, and 2070 respectively. This land cover degradation is more likely to occur in the proximity of roads, localities, highly populated areas, mining concessions, and high-altitudinal zones. Since the current land use system is unsustainable, there is a necessity to implement sustainable land management strategies that take into account the biophysical and socio-economic specificities of this region to reverse the problem of land cover degradation.

Machine learning insights on activities of daily living disorders in Chinese older adults

ObjectiveThis study on the aged population in China first used a large-scale longitudinal survey database to explore how different life factors affect their ability to engage in daily activities. We select and integrate multiple machine models to obtain an excellent model for analyzing relationships. Based on the identified factors, our goal is to help them maintain a good daily life and quality of life. MethodWe analyzed data from 13,220 older individuals participating in the China Longitudinal Health Longevity Survey (CLHLS) from 2002 to 2018. ADL was measured based on participants' self-reported results. Nine machine learning algorithms, including neural networks and an ensemble model, were employed with a 2/3 training and 1/3 testing split. Model performance was evaluated using the area under the curve (AUC), sensitivity, and specificity, while logistic regression assessed the relationship between lifestyle changes and ADL disorders. ResultThe K-nearest neighbors (KNN) and decision tree algorithms showed the best performance, with AUCs of 0.8598 and 0.8322, respectively. Combining results from all models improved the AUC to 0.8619. Activities, such as playing mahjong, engaging in outdoor work, and reducing TV time, were linked to lower ADL decline, with greater participation in social activities and pet care also being beneficial. ConclusionMachine learning algorithms, especially ensemble models, can effectively identify older adults at risk for ADL disorders. Increased outdoor activity, social engagement, and dietary adjustments are associated with a decreased risk of ADL deterioration. Translational significance1)The primary question addressed by this study is: What modifiable risk factors can impact Activities of Daily Living (ADL) in older adults?2)The main finding of this study is that specific daily activities, such as playing mahjong and engaging in outdoor activities, significantly reduce the risk of future ADL disorders in older adults. Additionally, a robust predictive model was developed using longitudinal data from 13,220 individuals, improving the accuracy of ADL disorder risk predictions.3)The meaning of the finding is that incorporating behavioral interventions into community care strategies can effectively enhance the well-being of older adults by minimizing their risk of ADL dysfunction.