Accuracy Of Prediction Model Research Articles

Development of a nomogram for predicting malnutrition in elderly hospitalized cancer patients: a cross-sectional study in China.

The Patient-Generated Subjective Global Assessment (PG-SGA) serves as a specialized nutritional assessment instrument designed for cancer patients. Despite its specificity, the complexity and time requirements of this tool, along with the necessity for administration by trained professionals, limit its practicality in clinical settings. Our objective is to identify a straightforward, efficient, and dependable nutritional assessment tool to promote broader adoption in clinical practice. This study encompassed a total of 450 patients diagnosed with cancer. Of these, 315 individuals constituted the training set, and the remaining 135 were allocated to the external validation set. The model variables were identified through the Least Absolute Shrinkage and Selection Operator (LASSO) regression method. Binary logistic regression outcomes facilitated the development of a nomogram, offering a visual depiction of the predicted probabilities. The predictive accuracy of the nomogram model was evaluated by calculating the area under the Receiver Operating Characteristic (ROC) curve. The LASSO method detected four variables that were included in the final prediction model: age, serum albumin levels (ALB), body mass index (BMI), and activities of daily living (ADL). The area under the curve (AUC) for this prediction model was 0.905. Both the internal and external calibration curves for malnutrition showed that the predictive nomogram model was highly accurate. The study has developed a prediction model that demonstrates remarkable accuracy in forecasting malnutrition. Furthermore, it presents a streamlined nutritional assessment tool aimed at swiftly identifying cancer patients at nutritional risk, thereby facilitating oncologists in delivering targeted nutritional support to these individuals.

Near miss prediction in commercial aviation through a combined model of grey neural network

Owing to unprecedented level of safety management, there have been few commercial aviation accidents recently, obstructing accurate prediction of safety trends. The innovative approach of proactive safety management was employed to replace reactive safety management for predicting commercial aviation near misses. In light of positive correlation between flight time and commercial aviation near-miss with different severity levels, this study aims to predict total / serious / non-serious near-miss per million flight hours. Considering grey system theory is good at prediction without adequate data, and artificial neural networks can handle nonlinear data well, the combined models of grey neural networks were developed for forecasting three data sequences over time respectively. Based on the empirical results of commercial aviation near-miss forecasting, BP neural network had the potential to enhance accuracy of grey prediction model. The three statistical measures of MAE/MSE/MAPE denoted the combined model of grey neural network outperformed the single model of GM (1, 1) or BP neural network. Predictions with a high degree of accuracy is beneficial to determining actual trends of commercial aviation near-misses at present or in future. Quantitative data will be offered for making specific decisions on maintaining preventive measures or promoting commercial aviation safety performance by introducing additional management actions.

A Comprehensive Review of Radio Signal Propagation Prediction for Terrestrial Wireless Communication Systems

The subject of study known as radio propagation prediction refers to predicting the behaviour and characteristics of radio waves as they propagate through the atmosphere. It is a basic element of all wireless communication systems, including satellite communications, broadcasting and cellular networks. While there is no study covers all the techniques used to predict the radio signal prediction, this study presents a review of the propagation prediction models between 2018-2023 used for terrestrial wireless communication systems; the classic empirical models were briefly explained, followed by the deterministic propagation models that have been developed using ray-tracing with deep and machine learning techniques. Recent studies on an improvement of the computational efficiency and accuracy of propagation prediction models were also reviewed, in addition to an overview of some of the traditional statistical models. Furthermore, some of the new techniques in propagation prediction were described. The results of these studies explain that techniques using ray tracing produce better results than other techniques.

A New Strategy: Remaining Useful Life Prediction of Wind Power Bearings Based on Deep Learning under Data Missing Conditions

With its formidable nonlinear mapping capabilities, deep learning has been widely applied in bearing remaining useful life (RUL) prediction. Given that equipment in actual work is subject to numerous disturbances, the collected data tends to exhibit random missing values. Furthermore, due to the dynamic nature of wind turbine environments, LSTM models relying on manually set parameters exhibit certain limitations. Considering these factors can lead to issues with the accuracy of predictive models when forecasting the remaining useful life (RUL) of wind turbine bearings. In light of this issue, a novel strategy for predicting the remaining life of wind turbine bearings under data scarcity conditions is proposed. Firstly, the average similarity (AS) is introduced to reconstruct the discriminator of the Generative Adversarial Imputation Nets (GAIN), and the adversarial process between the generative module and the discriminant is strengthened. Based on this, the dung beetle algorithm (DBO) is used to optimize multiple parameters of the long-term and short-term memory network (LSTM), and the complete data after filling is used as the input data of the optimized LSTM to realize the prediction of the remaining life of the wind power bearing. The effectiveness of the proposed method is verified by the full-life data test of bearings. The results show that, under the condition of missing data, the new strategy of AS-GAIN-LSTM is used to predict the RUL of wind turbine bearings, which has a more stable prediction performance.

Machine Learning-Based Design of Superhard High-Entropy Nitride Coatings.

Limited by the inefficiency of the conventional trial-and-error method and the boundless compositional design space of high-entropy alloys (HEAs), accelerating the discovery of superior-performing high-entropy nitride (HEN) coatings remains a formidable challenge. Herein, the superhard HEN coatings were designed and prepared using the rapidly developing data-driven model machine learning (ML). A database containing hardness and different features of HEN coatings was established and categorized into four subsets covering the information on composition, composition-physical descriptors, composition-technique parameters, and composition-physical descriptors-technique parameters. Feature engineering was employed to reduce dimensionality and interpret the impact of features on the evolution of hardness. Both root mean squared error (RMSE) and decision coefficient (R2) were applied to assess the predictive accuracy of ML models with different subsets, proportions of test set, and algorithms. The model with best predicted performance was used to explore superhard HEN coatings in a predefined virtual space. Among the generated 5-/6-/7-/8-component (excluding N) systems, the coating possessing highest hardness was individually selected for further preparation. Four newly prepared coatings achieved the superhard level with an average prediction error of 7.83%. The morphology, chemical composition, structure, and hardness of the newly prepared coatings were discussed. The nanocrystal-amorphous nanocomposite structure of the novel AlCrNbSiTiN coating with the highest hardness of 45.77 GPa was revealed. The results demonstrated that ML can effectively guide the design and composition optimization of superb-performance protective HEN coatings.

Comparison of model feature importance statistics to identify covariates that contribute most to model accuracy in prediction of insomnia.

Sleep is critical to a person's physical and mental health and there is a need to create high performing machine learning models and critically understand how models rank covariates. The study aimed to compare how different model metrics rank the importance of various covariates. A cross-sectional cohort study was conducted retrospectively using the National Health and Nutrition Examination Survey (NHANES), which is publicly available. This study employed univariate logistic models to filter out strong, independent covariates associated with sleep disorder outcome, which were then used in machine-learning models, of which, the most optimal was chosen. The machine-learning model was used to rank model covariates based on gain, cover, and frequency to identify risk factors for sleep disorder and feature importance was evaluated using both univariable and multivariable t-statistics. A correlation matrix was created to determine the similarity of the importance of variables ranked by different model metrics. The XGBoost model had the highest mean AUROC of 0.865 (SD = 0.010) with Accuracy of 0.762 (SD = 0.019), F1 of 0.875 (SD = 0.766), Sensitivity of 0.768 (SD = 0.023), Specificity of 0.782 (SD = 0.025), Positive Predictive Value of 0.806 (SD = 0.025), and Negative Predictive Value of 0.737 (SD = 0.034). The model metrics from the machine learning of gain and cover were strongly positively correlated with one another (r > 0.70). Model metrics from the multivariable model and univariable model were weakly negatively correlated with machine learning model metrics (R between -0.3 and 0). The ranking of important variables associated with sleep disorder in this cohort from the machine learning models were not related to those from regression models.

Improving Deep Learning-Based Algorithm for Ploidy Status Prediction Through Combined U-NET Blastocyst Segmentation and Sequential Time-Lapse Blastocysts Images.

Several approaches have been proposed to optimize the construction of an artificial intelligence-based model for assessing ploidy status. These encompass the investigation of algorithms, refining image segmentation techniques, and discerning essential patterns throughout embryonic development. The purpose of the current study was to evaluate the effectiveness of using U-NET architecture for embryo segmentation and time-lapse embryo image sequence extraction, three and ten hr before biopsy to improve model accuracy for prediction of embryonic ploidy status. A total of 1.020 time-lapse videos of blastocysts with known ploidy status were used to construct a convolutional neural network (CNN)-based model for ploidy detection. Sequential images of each blastocyst were extracted from the time-lapse videos over a period of three and ten hr prior to the biopsy, generating 31.642 and 99.324 blastocyst images, respectively. U-NET architecture was applied for blastocyst image segmentation before its implementation in CNN-based model development. The accuracy of ploidy prediction model without applying the U-NET segmented sequential embryo images was 0.59 and 0.63 over a period of three and ten hr before biopsy, respectively. Improved model accuracy of 0.61 and 0.66 was achieved, respectively with the implementation of U-NET architecture for embryo segmentation on the current model. Extracting blastocyst images over a 10 hr period yields higher accuracy compared to a three-hr extraction period prior to biopsy. Combined implementation of U-NET architecture for blastocyst image segmentation and the sequential compilation of ten hr of time-lapse blastocyst images could yield a CNN-based model with improved accuracy in predicting ploidy status.

Deciphering the role of cuproptosis-related lncRNAs in shaping the lung cancer immune microenvironment: A comprehensive prognostic model.

Cuproptosis plays an important role in cancer, but its role in lung cancer remains unknown. Transcriptional profiles, clinical details and mutation data were acquired from the Cancer Genome Atlas database through a variety of methods. The analysis of this publicly available data was comprehensively performed using R software along with its relevant packages, ensuring a thorough examination of the information. In this study, we conducted a detailed analysis of cuproptosis-related genes and lncRNA co-expression, identifying 129 relevant lncRNAs and establishing a prognostic model with four key lncRNAs (LINC00996, RPARP-AS1, SND1-IT1, TMPO-AS1). Utilizing data from TCGA and GEO databases, the model effectively categorized patients into high- and low-risk groups, showing significant survival differences. Correlation analysis highlighted specific relationships between individual lncRNAs and cuproptosis genes. Our survival analysis indicated a higher survival rate in the low-risk group across various cohorts. Additionally, the model's predictive accuracy was confirmed through independent prognostic analysis and ROC curve evaluations. Functional enrichment analysis revealed distinct biological pathways and immune functions between risk groups. Tumour mutation load analysis differentiated high- and low-risk groups by their mutation profiles. Drug sensitivity analysis and immune infiltration studies using the CIBERSORT algorithm further elucidated the potential treatment responses in different risk groups. This comprehensive evaluation underscores the significance of lncRNAs in cuproptosis and their potential as biomarkers for lung cancer prognosis and immune microenvironment.

Prognostic model development and clinical correlation of eight key genes in skin cutaneous melanoma

Cutaneous melanoma (SKCM) is a challenging and increasingly prevalent cancer with limited effective treatments. In our extensive study of 342 SKCM samples, we developed a prognostic model identifying eight key genes—CASPASE7CLEAVEDD198, FOXO3A, Melanoma gp100, CD171, 1433ZETA, SRC, P21, and CABL—linked to SKCM prognosis. Statistical analysis indicated significant differences in clinical outcomes between low and high-risk groups, corroborated by principal component analysis (PCA). Survival analysis and receiver operating characteristic (ROC) curve analysis confirmed the model's predictive accuracy for SKCM prognosis. Additionally, we observed notable correlations between the expression levels of genes related to prognosis and clinical characteristics. Our research offers crucial insights into SKCM prognosis, suggesting potential diagnostic markers and personalized treatment targets.

Brain-age prediction: Systematic evaluation of site effects, and sample age range and size.

Structural neuroimaging data have been used to compute an estimate of the biological age of the brain (brain-age) which has been associated with other biologically and behaviorally meaningful measures of brain development and aging. The ongoing research interest in brain-age has highlighted the need for robust and publicly available brain-age models pre-trained on data from large samples of healthy individuals. To address this need we have previously released a developmental brain-age model. Here we expand this work to develop, empirically validate, and disseminate a pre-trained brain-age model to cover most of the human lifespan. To achieve this, we selected the best-performing model after systematically examining the impact of seven site harmonization strategies, age range, and sample size on brain-age prediction in a discovery sample of brain morphometric measures from 35,683 healthy individuals (age range: 5-90 years; 53.59% female). The pre-trained models were tested for cross-dataset generalizability in an independent sample comprising 2101 healthy individuals (age range: 8-80 years; 55.35% female) and for longitudinal consistency in a further sample comprising 377 healthy individuals (age range: 9-25 years; 49.87% female). This empirical examination yielded the following findings: (1) the accuracy of age prediction from morphometry data was higher when no site harmonization was applied; (2) dividing the discovery sample into two age-bins (5-40 and 40-90 years) provided a better balance between model accuracy and explained age variance than other alternatives; (3) model accuracy for brain-age prediction plateaued at a sample size exceeding 1600 participants. These findings have been incorporated into CentileBrain (https://centilebrain.org/#/brainAGE2), an open-science, web-based platform for individualized neuroimaging metrics.

Eleven inflammation-related genes risk signature model predicts prognosis of patients with breast cancer.

Changes in gene expression are associated with malignancy. Analysis of gene expression data could be used to reveal cancer subtypes, key molecular drivers, and prognostic characteristics and to predict cancer susceptibility, treatment response, and mortality. It has been reported that inflammation plays an important role in the occurrence and development of tumors. Our aim was to establish a risk signature model of breast cancer with inflammation-related genes (IRGs) to evaluate their survival prognosis. We downloaded 200 IRGs from the Molecular Signatures Database (MSigDB). The data of breast cancer were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO). Differential gene expression analysis, the least absolute shrinkage and selection operator (LASSO), Cox regression analysis, and overall survival (OS) analysis were used to construct a multiple-IRG risk signature. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were carried out to annotate functions of the differentially expressed IRGs (DEIRGs) The predictive accuracy of the prognostic model was evaluated by time-dependent receiver operating characteristic (ROC) curves. Subsequently, nomograms were constructed to guide clinical application according to the univariate and multivariate Cox proportional hazards regression analyses. Eventually, we applied gene set variation analysis (GSVA), mutation analysis, immune infiltration analysis, and drug response analysis to compare the differences between high- and low-risk patients. Totally, 65 DEIRGs were obtained after comparing 1,092 breast cancer tissues with 113 paracancerous tissues in TCGA. Among them, 11 IRGs (IL18, IL12B, RASGRP1, HPN, CLEC5A, SCARF1, TACR3, VIP, CCL2, CALCRL, ABCA1) were screened with nonzero coefficient by LASSO regression analysis to construct the prognostic model, which was validated in GSE96058.The 11-gene IRGs risk signature model stratified patients into high- or low-risk groups, with those in the low-risk group having longer survival time and less deaths. Multivariate Cox analysis manifested that risk score, age, and stage were the three independent prognostic factors for breast cancer patients. There were 12 pathways with higher activities and 24 pathways with lower activities in the high-risk group compared with the low-risk group, yet no difference of gene mutation load was observed between the two groups. In immune infiltration analysis, we noted that the proportion of T cells showed a decreased trend according to the increase of risk score and most of the immune cells were enriched in the low-risk group. Inversely, macrophages M2 were more highly distributed in the high-risk group. We identified 67 approved drugs that showed a different effect between the high- and low-risk patients and the top 2 gene-drug pairs were IL12B-sunitinib and SCARF1-ruxolitinib. The 11-IRG risk signature model is a promising tool to predict the survival of breast cancer patients and the expressions of IL12B and SCARF1 may serve as potential targets for therapy of breast cancer.

A comparison of models for predicting the maximum spreading factor in droplet impingement

The maximum spreading factor during droplet impact on a dry surface is a pivotal parameter of a range of applications, including inkjet printing, anti-icing, and micro-droplet transportation. It is determined by a combination of the inertial force, viscous force, surface tension, and fluid–solid interaction. There are currently a series of qualitative and quantitative prediction models for the maximum spreading factor rooted in both momentum and energy conservation. However, the performance of these models on consistent experimental samples remains ambiguous. In this work, a comprehensive set of 785 experimental samples spanning the last four decades is compiled. These samples encompass Weber numbers ranging from 0.038 to 2447.7 and Reynolds numbers from 9 to 34 339. A prediction model is introduced that employs a neural network, which achieves an average relative error of less than 16.6% with a standard error of 0.018 08 when applied to the test set. Following this, a fair comparison is presented of the accuracy, generality, and stability of different prediction models. Although the neural network model provides superior accuracy and generality, its stability is weaker than that of Scheller's We-Re-dependent formula, chiefly due to the absence of physical constraints. Subsequently, a physics-informed prediction model is introduced by considering a physical loss term. This model demonstrates comprehensive enhancements compared to the original neural network, and the average relative and standard errors for this model are reduce to 13.6% and 0.010 59, respectively. This novel model should allow for the rapid and precise prediction of the maximum spreading factor across a broad range of parameters for various applications.

INTEGRATING ARTIFICIAL INTELLIGENCE AND MATHEMATICAL MODELS FOR PREDICTIVE MAINTENANCE IN INDUSTRIAL SYSTEMS

Predictive maintenance is a critical task for ensuring the reliability and efficiency of industrial systems. The integration of artificial intelligence (AI) and mathematical models has shown great potential in improving the accuracy and efficiency of predictive maintenance. This study provides an overview of the different types of mathematical models used for predictive maintenance, including physics-based, data-driven, and hybrid models. The study also discusses how AI techniques, such as machine learning and deep learning, can be used to enhance the accuracy and efficiency of predictive maintenance models. Additionally, the article highlights some of the challenges and limitations of integrating AI and mathematical models for predictive maintenance in industrial systems. Finally, this study provides a case study to demonstrate the practical application of the integrated approach for predictive maintenance in an industrial setting. This article aims to provide a comprehensive overview of the state-of-the-art in integrating AI and mathematical models for predictive maintenance and to provide guidance for researchers and practitioners working in this field.

A Hybrid Prediction Model for Rock Reservoir Bank Slope Deformation Considering Fractured Rock Mass Parameters

Deformation monitoring data provide a direct representation of the structural behavior of reservoir bank rock slopes, and accurate deformation prediction is pivotal for slope safety monitoring and disaster warning. Among various deformation prediction models, hybrid models that integrate field monitoring data and numerical simulations stand out due to their well-defined physical and mechanical concepts, and their ability to make effective predictions with limited monitoring data. The predictive accuracy of hybrid models is closely tied to the precise determination of rock mass mechanical parameters in structural numerical simulations. However, rock masses in rock slopes are characterized by intersecting geological structural planes, resulting in reduced strength and the creation of multiple fracture flow channels. These factors contribute to the heterogeneous, anisotropic, and size-dependent properties of the macroscopic deformation parameters of the rock mass, influenced by the coupling of seepage and stress. To improve the predictive accuracy of the hybrid model, this study introduces the theory of equivalent continuous media. It proposes a method for determining the equivalent deformation parameters of fractured rock mass considering the coupling of seepage and stress. This method, based on a discrete fracture network (DFN) model, is integrated into the hybrid prediction model for rock slope deformation. Engineering case studies demonstrate that this approach achieves a high level of prediction accuracy and holds significant practical value.

Design and Development of Artificial Intelligence Based Applications Using Classification Algorithms in Datamining to Predict Early Diabetes

Diabetes is a chronic disease characterized by high blood sugar levels. High blood sugar levels can cause various serious complications, such as heart disease, stroke and kidney failure. Early detection of diabetes is essential to prevent this complication. This research is using experimental method. The experimental method is a research method carried out by deliberately manipulating the independent variable to see its effect on the dependent variable. In this case, the independent variable being manipulated is the machine learning algorithm used to predict diabetes. The dependent variable observed was the accuracy of the diabetes prediction model. From the research results, it was found that the Naive Bayes model had the highest accuracy, namely 81.5%. This model is better than the Decision Tree C.45 (72.7%), Random Forest (73.8%), and K-Nearest Neighbor (71.5%) models. Based on the research results, it can be concluded that the Naive Bayes model is the best model for predicting diabetes. The appropriate system development method for this research is the Rapid Application Development (RAD) method. Further research can be carried out to improve the accuracy of diabetes prediction models by using optimization models, such as Particle Swarm Optimization or Colony Optimization. In addition, further research can be carried out to develop diabetes prediction applications that can predict diseases that may be caused by diabetes, such as heart disease, stroke, kidney disease and blindness.

Towards Enhanced Autonomous Driving Takeovers: Fuzzy Logic Perspective for Predicting Situational Awareness

This paper investigates the application of fuzzy logic to enhance situational awareness in Advanced Driver Assistance Systems (ADAS). Situational awareness is critical for drivers to respond appropriately to dynamic driving scenarios. As car automation increases, monitoring situational awareness ensures that drivers can effectively take control of the vehicle when needed. Our study explores whether fuzzy logic can accurately assess situational awareness using a set of 14 critical predictors categorized into time decision, criticality, eye-related metrics, and driver experience. We based our work on prior research that used machine learning (ML) models to achieve high accuracy. Our proposed fuzzy logic system aims to match the predictive accuracy of ML models while providing additional benefits in terms of interpretability and robustness. This approach emphasizes a fresh perspective on situational awareness within ADAS, potentially improving safety and efficiency in real-world driving scenarios.

Study on the Cutting Performance and Remaining Life Prediction of Micro-Texture Ball End Milling Cutters for Titanium Alloys

As a fundamental machining tool, the ball end milling cutter plays a crucial role in manufacturing. Due to its low thermal conductivity, the heat generated during the cutting process of titanium alloy materials is not dissipated efficiently, resulting in a substantial cutting heat. This heat leads to chip adhesion and exacerbates the wear of the ball end milling cutter, ultimately affecting its service life. Therefore, studying the residual life of the tool during the cutting process is essential to prevent significant impacts on the product’s surface quality due to tool damage and passivation. Most research on micro-texture cutters is based on experiments that analyze the wear patterns of cutters under various lubrication conditions and their influence on the cutting process. Different neural network prediction models are employed to enhance the accuracy and stability of tool life prediction models. However, the exploration of other superior models for predicting the life of micro-texture cutters remains ongoing. This paper is based on an experiment involving the milling of titanium alloy using a micro-pit-structured ball end milling cutter. It was found that the cutting force of the tool is higher during the initial and later wear stages. During the stable wear stage, the unevenness of the defective layer on the tool surface is reduced, increasing the contact area and reducing the surface pressure, thereby decreasing the cutting force. This study analyzes the influence of micro-pit structural parameters on the wear and milling force of the ball end milling cutter. By evaluating the wear value of the ball end milling cutter after each cut, the wear mechanism of the micro-texture cutter is identified. A deep-learning-based bidirectional long short-term memory (BiLSTM) neural network model for tool life prediction is developed. Through training and validation, the model’s accuracy and stability are continuously improved. A comparative analysis with different predictive models is conducted to determine whether the proposed model offers advantages over existing models, which is crucial for maximizing tool utilization and reducing manufacturing costs.

Prediction of ovarian cancer using artificial intelligence tools.

Ovarian cancer is a common type of cancer and a leading cause of death in women. Therefore, accurate and fast prediction of ovarian tumors is crucial. One of the appropriate and precise methods for predicting and diagnosing this cancer is to build a model based on artificial intelligence methods. These methods provide a tool for predicting ovarian cancer according to the characteristics and conditions of each person. In this study, a data set included records related to 171 cases of benign ovarian tumors, and 178 records related to cases of ovarian cancer were analyzed. The data set contains the records of blood test results and tumor markers of the patients. After data preprocessing, including removing outliers and replacing missing values, the weight of the effective factors was determined using information gain indices and the Gini index. In the next step, predictive models were created using random forest (RF), support vector machine (SVM), decision trees (DT), and artificial neural network (ANN) models. The performance of these models was evaluated using the 10-fold cross-validation method using the indicators of specificity, sensitivity, accuracy, and the area under the receiver operating characteristic curve. Finally, by comparing the performance of the models, the best predictive model of ovarian cancer was selected. The most important predictive factors were HE4, CA125, and NEU. The RF model was identified as the best predictive model, with an accuracy of more than 86%. The predictive accuracy of DT, SVM, and ANN models was estimated as 82.91%, 85.25%, and 79.35%, respectively. Various artificial intelligence (AI) tools can be used with high accuracy and sensitivity in predicting ovarian cancer. Therefore, the use of these tools can help specialists and patients with early, easier, and less expensive diagnosis of ovarian cancer. Future studies can leverage AI to integrate image data with serum biomarkers, thereby facilitating the creation of novel models and advancing the diagnosis and treatment of ovarian cancer.

Prediction of Geometric Dimensions of Deposited Layer Produced Using Laser-Arc Hybrid Additive Manufacturing.

Laser-arc hybrid additive manufacturing (LAHAM) holds substantial potential in industrial applications, yet ensuring dimensional accuracy remains a major challenge. Accurate prediction and effective control of the geometrical dimensions of the deposited layers are crucial for achieving this accuracy. The width and height of the deposited layers, key indicators of geometric dimensions, directly affect the forming precision. This study conducted experiments and in-depth analysis to investigate the influence of various process parameters on these dimensions and proposed a predictive model for accurate forecasting. It was found that the width of the deposited layers was positively correlated with laser power and arc current and negatively correlated with scanning speed, while the height was negatively correlated with laser power and scanning speed and positively with arc current. Quantitative analysis using the Taguchi method revealed that the arc current had the most significant impact on the dimensions of the deposited layers, followed by scanning speed, with laser power having the least effect. A predictive model based on extreme gradient boosting (XGBoost) was developed and optimized using particle swarm optimization (PSO) for tuning the number of leaf nodes, learning rate, and regularization coefficients, resulting in the PSO-XGBoost model. Compared to models enhanced with PSO-optimized support vector regression (SVR) and XGBoost, the PSO-XGBoost model exhibited higher accuracy, the smallest relative error, and performed better in terms of Mean Relative Error (MRE), Mean Square Error (MSE), and Coefficient of Determination R2 metrics. The high predictive accuracy and minimal error variability of the PSO-XGBoost model demonstrate its effectiveness in capturing the complex nonlinear relationships between process parameters and layer dimensions. This study provides valuable insights for controlling the geometric dimensions of the deposited layers in LAHAM.

Ensemble learning soft sensor method of endpoint carbon content and temperature of BOF based on GCN embedding supervised ensemble clustering

Abstract The endpoint control of Basic Oxygen Furnace (BOF) steelmaking depends on the prediction of the endpoint carbon content and temperature. However, predicting these variables is challenging because of the numerous working conditions in the industrial field and the volatility of the sensor data collected during BOF steelmaking. The accuracy of prediction models in ensemble learning depends significantly on the initial distribution of data. However, the complex nature of BOF steelmaking data makes it challenging to generate diverse subsets, which ultimately affects the accuracy of predictions. This paper presents a new approach called Graph Convolutional Network Node Embedding Supervised Ensemble Clustering (GESupEC) for soft sensor modelling in ensemble learning to tackle these issues. GESupEC utilizes a similarity graph derived from a co-association matrix and employs graph convolutional networks to extract structural information among nodes. By optimising the clustering loss within the network, GESupEC learns compact node representations that are useful for the clustering task. Furthermore, it generates a reconstruction matrix based on the similarity of node embeddings. This matrix helps with the extraction of a suitable subset of data for BOF steelmaking through matrix decomposition. After that, the gradient boosting decision tree regression sub-model is established based on the data subset. An ensemble strategy called Gray Relational Analysis Weighted Average is proposed, which assigns weights based on the grey relation similarity between test samples and different data subsets. This weighted average strategy aims to enhance the accuracy of carbon content and temperature predictions. When tested with actual BOF steelmaking generation process data, the prediction accuracy of carbon content reached 88.6% within the error range of ±0.02%, and the prediction accuracy of temperature reached 92.6% within the error range of ±10 °C.