Multivariate Prediction Research Articles

Echocardiographic predictors of cardiovascular outcome in heart failure with preserved ejection fraction

Abstract Background The optimal echocardiographic predictors of cardiovascular outcome in heart failure with presereved ejection fraction (HFpEF) are unknown. Purpose We aimed to identify independent echocardiographic predictors of cardiovascular outcome in patients with HFpEF. Methods Systematic literature search of three electronic databases was conducted from date of inception until November 2022. Hazard ratios (HRs) and their 95% confidence intervals (CIs) for echocardiographic variables from multivariate prediction models for the composite primary endpoint of cardiovascular death and heart failure (HF) hospitalization were pooled using a random effects meta-analysis. Specific subgroup analyses were conducted for studies that enrolled patients with acute versus chronic HF, and for those studies that included E/e’, pulmonary artery systolic pressure (PASP), renal function, natriuretic peptides and diuretic use in multivariate models. Results Forty-six studies totalling 20,056 patients with HFpEF were included. Three echocardiographic parameters emerged as independent predictors in all subgroup analyses: Decreased left ventricular (LV) global longitudinal strain (HR 1.24, 95% CI 1.10-1.39 per 5% decrease), decreased left atrial (LA) reservoir strain (HR 1.30, 95% CI 1.13-1.1.50 per 5% decrease) and lower tricuspid annular plane systolic excursion (TAPSE) to PASP ratio (HR 1.17, 95% CI 1.07-1.25) per 0.1 unit decrease). Other independent echocardiographic predictors of the primary endpoint were a higher E/e’, moderate to severe tricuspid regurgitation, LV mass index and LA ejection fraction, although these variables were less robust. Conclusion Impaired LV global longitudinal strain, lower LA reservoir strain and lower TAPSE/PASP ratio predict cardiovascular death and HF hospitalization in HFpEF and are independent of filling pressures, clinical characteristics and natriuretic peptides. These echocardiographic parameters reflect key functional changes in HFpEF, and should be incorporated in future prospective risk prediction models.

Analysis of the efficacy of Prostatic Artery Embolization in the treatment of Benign Prostatic Hyperplasia.

To investigate the safety and efficacy of prostatic artery embolization (PAE) in the treatment of lower urinary tract symptoms (LUTS) caused by benign prostatic hyperplasia (BPH), and to investigate predictors of clinical success of PAE. A retrospective analysis was used to collect 107 patients [median age 81.0 (73.0,85.0)] with BPH-related LUTS treated with PAE from September 2014 to February 2022 in a hospital. Repeated measurement ANOVA was used to compare the efficacy evaluation indicators at different times before and after PAE. Univariate and multivariate analyses were used to identify potential predictors of PAE clinical success and establish the optimal joint prediction model. The Receiver Operating Characteristic curves of the quantitative predictors and multivariate model prediction probability values significantly correlated with clinical success were plotted. Of the 107 cases, 103 (96.3%) successfully underwent PAE. The International Prostate Symptom Score (IPSS) decreased from a baseline mean of 24.94 to 10.19 (P < 0.05) 3 months after PAE, and the mean IPSS at 6 months, 12 months and 24 months was 10.12, 11.30 and 11.86, respectively, which were statistically significant compared with baseline (P<0.05). Predictors of clinical success were greater prostate volume (> 65 ml, P = 0.018), adenomatous-dominant benign prostatic hyperplasia (AdBPH)(P = 0.030), indwelling catheterization due to urine retention (P = 0.028), and bilateral embolization (P = 0.018). PAE was able to significantly improve BPH-related LUTS and the outcome indicators remained stable at long-term follow-up. Preoperative urinary retention catheters, AdBPH, larger prostate volume (> 65ml) and bilateral embolization suggest better clinical efficacy.

Subclassification of obesity for precision prediction of cardiometabolic diseases.

Obesity and cardiometabolic disease often, but not always, coincide. Distinguishing subpopulations within which cardiometabolic risk diverges from the risk expected for a given body mass index (BMI) may facilitate precision prevention of cardiometabolic diseases. Accordingly, we performed unsupervised clustering in four European population-based cohorts (N ≈ 173,000). We detected five discordant profiles consisting of individuals with cardiometabolic biomarkers higher or lower than expected given their BMI, which generally increases disease risk, in total representing ~20% of the total population. Persons with discordant profiles differed from concordant individuals in prevalence and future risk of major adverse cardiovascular events (MACE) and type 2 diabetes. Subtle BMI-discordances in biomarkers affected disease risk. For instance, a 10% higher probability of having a discordant lipid profile was associated with a 5% higher risk of MACE (hazard ratio in women 1.05, 95% confidence interval 1.03, 1.06, P = 4.19 × 10-10; hazard ratio in men 1.05, 95% confidence interval 1.04, 1.06, P = 9.33 × 10-14). Multivariate prediction models for MACE and type 2 diabetes performed better when incorporating discordant profile information (likelihood ratio test P < 0.001). This enhancement represents an additional net benefit of 4-15 additional correct interventions and 37-135 additional unnecessary interventions correctly avoided for every 10,000 individuals tested.

Research on a Non-Stationary Groundwater Level Prediction Model Based on VMD-iTransformer and Its Application in Sustainable Water Resource Management of Ecological Reserves

The precise forecasting of groundwater levels significantly influences plant growth and the sustainable management of ecosystems. Nonetheless, the non-stationary characteristics of groundwater level data often hinder the current deep learning algorithms from precisely capturing variations in groundwater levels. We used Variational Mode Decomposition (VMD) and an enhanced Transformer model to address this issue. Our objective was to develop a deep learning model called VMD-iTransformer, which aims to forecast variations in the groundwater level. This research used nine groundwater level monitoring stations located in Hangjinqi Ecological Reserve in Kubuqi Desert, China, as case studies to forecast the groundwater level over four months. To enhance the predictive performance of VMD-iTransformer, we introduced a novel approach to model the fluctuations in groundwater levels in the Kubuqi Desert region. This technique aims to achieve precise predictions of the non-stationary groundwater level conditions. Compared with the classic Transformer model, our deep learning model more effectively captured the non-stationarity of groundwater level variations and enhanced the prediction accuracy by 70% in the test set. The novelty of this deep learning model lies in its initial decomposition of multimodal signals using an adaptive approach, followed by the reconfiguration of the conventional Transformer model’s structure (via self-attention and inversion of a feed-forward neural network (FNN)) to effectively address the challenge of multivariate time prediction. Through the evaluation of the prediction results, we determined that the method had a mean absolute error (MAE) of 0.0251, a root mean square error (RMSE) of 0.0262, a mean absolute percentage error (MAPE) of 1.2811%, and a coefficient of determination (R2) of 0.9287. This study validated VMD and the iTransformer deep learning model, offering a novel modeling approach for precisely predicting fluctuations in groundwater levels in a non-stationary context, thereby aiding sustainable water resource management in ecological reserves. The VMD-iTransformer model enhances projections of the water level, facilitating the reasonable distribution of water resources and the long-term preservation of ecosystems, providing technical assistance for ecosystems’ vitality and sustainable regional development.

Development and construction of a cataract risk prediction model based on biochemical indices: the National Health and Nutrition Examination Survey, 2005–2008

PurposeThe aim of this study is to develop and validate a novel multivariable prediction model capable of accurately estimating the probability of cataract development, utilizing parameters such as blood biochemical markers and age.DesignThis population-based cross-sectional study comprised 9,566 participants drawn from the National Health and Nutrition Examination Survey (NHANES) across the 2005–2008 cycles.MethodsDemographic information and laboratory test results from the patients were collected and analyzed using LASSO regression and multivariate logistic regression to accurately capture the influence of biochemical indicators on the outcomes. The SHAP (Shapley Additive Explanations) scale was employed to assess the importance of each clinical feature, excluding age. A multivariate logistic regression model was then developed and visualized as a nomogram. To assess the model’s performance, its discrimination, calibration, and clinical utility were evaluated using receiver operating characteristic (ROC) curves, 10-fold cross-validation, Hosmer-Lemeshow calibration curves, and decision curve analysis (DCA), respectively.ResultsLogistic regression analysis identified age, erythrocyte folate (nmol/L), blood glucose (mmol/L), and blood urea nitrogen (mmol/L) as independent risk factors for cataract, and these variables were incorporated into a multivariate logistic regression-based nomogram for cataract risk prediction. The area under the receiver operating characteristic (ROC) curve (AUC) for cataract risk prediction was 0.917 (95% CI: 0.9067–0.9273) in the training cohort, and 0.9148 (95% CI: 0.8979–0.9316) in the validation cohort. The Hosmer-Lemeshow calibration curve demonstrated a good fit, indicating strong model calibration. Ten-fold cross-validation confirmed the logistic regression model’s robust predictive performance and stability during internal validation. Decision curve analysis (DCA) demonstrated that the nomogram prediction model provided greater clinical benefit for predicting cataract risk when the patient’s threshold probability ranged from 0.10 to 0.90.ConclusionThis study identified blood urea nitrogen (mmol/L), serum glucose (mmol/L), and erythrocyte folate (mmol/L) as significant risk factors for cataract. A risk prediction model was developed, demonstrating strong predictive accuracy and clinical utility, offering clinicians a reliable tool for early and effective diagnosis. Cataract development may be delayed by reducing levels of blood urea nitrogen, serum glucose, and erythrocyte folate through lifestyle improvements and dietary modifications.

Multiple Load Forecasting of Integrated Renewable Energy System Based on TCN-FECAM-Informer

In order to solve the problem of complex coupling characteristics between multivariate load sequences and the difficulty in accurate multiple load forecasting for integrated renewable energy systems (IRESs), which include low-carbon emission renewable energy sources, in this paper, the TCN-FECAM-Informer multivariate load forecasting model is proposed. First, the maximum information coefficient (MIC) is used to correlate the multivariate loads with the weather factors to filter the appropriate features. Then, effective information of the screened features is extracted and the frequency sequence is constructed using the frequency-enhanced channel attention mechanism (FECAM)-improved temporal convolutional network (TCN). Finally, the processed feature sequences are sent to the Informer network for multivariate load forecasting. Experiments are conducted with measured load data from the IRES of Arizona State University, and the experimental results show that the TCN and FECAM can greatly improve the multivariate load prediction accuracy and, at the same time, demonstrate the superiority of the Informer network, which is dominated by the attentional mechanism, compared with recurrent neural networks in multivariate load prediction.

Recommendations for prediction models in clinical practice guidelines for cardiovascular diseases are over-optimistic: a global survey utilizing a systematic literature search

BackgroundThis study aimed to synthesize the recommendations for prediction models in cardiovascular clinical practice guidelines (CPGs) and assess the methodological quality of the relevant primary modeling studies.MethodsWe performed a systematic literature search of all available cardiovascular CPGs published between 2018 and 2023 that presented specific recommendations (whether in support or non-support) for at least one multivariable clinical prediction model. For the guideline-recommended models, the assessment of the methodological quality of their primary modeling studies was conducted using the Prediction model Risk Of Bias ASsessment Tool (PROBAST).ResultsIn total, 46 qualified cardiovascular CPGs were included, with 69 prediction models and 80 specific recommendations. Of the 80 specific recommendations, 74 supported 57 models (53 were fully recommended and 4 were conditionally recommended) in cardiovascular practice with moderate to strong strength. Most of the guideline-recommended models were focused on predicting prognosis outcomes (53/57, 93%) in primary and tertiary prevention, focusing primarily on long-term risk stratification and prognosis management. A total of 10 conditions and 7 types of target population were involved in the 57 models, while heart failure (14/57, 25%) and a general population with or without cardiovascular risk factor(s) (12/57, 21%) received the most attention from the guidelines. The assessment of the methodological quality of 57 primary studies on the development of the guideline-recommended models revealed that only 40% of the modeling studies had a low risk of bias (ROB). The causes of high ROB were mainly in the analysis and participant domains.ConclusionsGlobal cardiovascular CPGs presented an unduly positive appraisal of the existing prediction models in terms of ROB, leading to stronger recommendations than were warranted. Future cardiovascular practice may benefit from well-established clinical prediction models with better methodological quality and extensive external validation.

Multivariable High-Dimension Time-Series Prediction in SIoT via Adaptive Dual-Graph-Attention Encoder-Decoder With Global Bayesian Optimization

Multivariable High-Dimension Time-Series Prediction in SIoT via Adaptive Dual-Graph-Attention Encoder-Decoder With Global Bayesian Optimization

Human serum albumin: prediction model and reference values for preterm and term neonates.

Human serum albumin (HSA) concentrations may alter HSA-bound drug distribution. This study aims to describe longitudinal real-world HSA trends, and to develop a prediction model for HSA concentrations using a large neonatal cohort. Patients admitted to the neonatal intensive care unit of the University Hospitals Leuven (postnatal age (PNA) ≤28days) were retrospectively included. Using linear mixed models, covariate effects on HSA were explored. A multivariable prediction model was developed (backward model selection procedure, 1% significance level). In total, 848 neonates were included [median(interquartile range) gestational age (GA) 35(32-38)weeks, birth weight (BW) 2400(1640-3130)grams]. Median HSA concentration was 32.3(28.7-35.6)g/L. Longitudinal analyses demonstrated increasing HSA concentrations with PNA and GA for most GA groups. Univariable analyses revealed significant associations of HSA with PNA, GA, BW, current weight, total and direct bilirubin, total plasma proteins, respiratory support, mechanical ventilation, sepsis, ibuprofen use, and C-reactive protein (p-values < 0.05). A high-performance (R2 = 76.3%) multivariable HSA prediction model was developed, and PNA- and GA-dependent HSA centiles were provided. Population-specific HSA centiles and an accurate neonatal HSA prediction model were developed, incorporating both maturational and non-maturational covariates. These results can enhance future clinical care and pharmacokinetic analyses to improve pharmacotherapy of HSA-bound drugs in neonates, respectively. To improve future pharmacokinetic modeling initiatives, a high-performance human serum albumin (HSA) prediction model was developed for (pre)term neonates, using a large, single-center cohort of real-world data. This prediction model integrates both maturational and non-maturational covariates, resulting in accurate HSA predictions in neonates. Additionally, HSA centiles based on postnatal and gestational age were developed, which can be easily applied in clinical practice when interpreting HSA concentrations of neonates. In general, unbound drug fractions are higher in neonates compared to older populations. To improve pharmacotherapy of HSA-bound drugs in neonates, the obtained results can be integrated in future pharmacokinetic-pharmacodynamic analyses.

Multivariable prediction of functional outcome after first-episode psychosis: a crossover validation approach in EUFEST and PSYSCAN.

Several multivariate prognostic models have been published to predict outcomes in patients with first episode psychosis (FEP), but it remains unclear whether those predictions generalize to independent populations. Using a subset of demographic and clinical baseline predictors, we aimed to develop and externally validate different models predicting functional outcome after a FEP in the context of a schizophrenia-spectrum disorder (FES), based on a previously published cross-validation and machine learning pipeline. A crossover validation approach was adopted in two large, international cohorts (EUFEST, n = 338, and the PSYSCAN FES cohort, n = 226). Scores on the Global Assessment of Functioning scale (GAF) at 12 month follow-up were dichotomized to differentiate between poor (GAF current < 65) and good outcome (GAF current ≥ 65). Pooled non-linear support vector machine (SVM) classifiers trained on the separate cohorts identified patients with a poor outcome with cross-validated balanced accuracies (BAC) of 65-66%, but BAC dropped substantially when the models were applied to patients from a different FES cohort (BAC = 50-56%). A leave-site-out analysis on the merged sample yielded better performance (BAC = 72%), highlighting the effect of combining data from different study designs to overcome calibration issues and improve model transportability. In conclusion, our results indicate that validation of prediction models in an independent sample is essential in assessing the true value of the model. Future external validation studies, as well as attempts to harmonize data collection across studies, are recommended.

FAFTransformer: Multivariate time series prediction method based on multi‐period feature recombination

AbstractMultivariate time series forecasting is widely used in various fields in real life. Many time series prediction models have been proposed. The current forecasting model lacks the mining of correlation between sequences based on different periods and correlation of periodical features between different periods when dealing with data. In this paper, we propose a multivariate data prediction model FAFTransformer based on the reorganization of multi‐periodic features, which first extracts the multi‐periodic information of the time series using the method of frequency domain analysis. The temporal dependencies within sequences are then captured using convolution based on different periods, and the correlations between sequences are learned by combining the multivariate attention mechanism to obtain the intra‐sequence and inter‐sequence correlations under the same period. Finally, period fusion is proposed to capture the correlation of period characteristics between different periods. The experimental results show that the model achieves the best results on multiple datasets compared to the latest seven predictive models for time‐series data.

Predictive Value of CCQ and its Subdomains for Hospitalized Exacerbations in Chronic Obstructive Pulmonary Disease: A 3-Year Observational Prospective Study.

The prediction capacity of the Clinical COPD Questionnaire (CCQ) and its functional, symptom, and mental subdomain for COPD hospitalized exacerbation were rarely studied. To examine the prognostic capacity of the total CCQ and its three subdomains for 3-year COPD hospitalized exacerbations. We analyzed the predictive ability of total CCQ score and its subdomains for hospitalized exacerbations in an observational cohort of 987 subjects with stable COPD from the RealDTC, an ongoing multicenter prospective study. Hospitalized exacerbations were prospectively collected every 6month for a maximum of 3years. The total CCQ and its functional and symptom domain, but not the mental domain, were significantly associated with 3-year hospitalized exacerbations by multivariate Cox regression analysis. The predictive capacity of functional domain was similar to that of the total CCQ, but significantly stronger than the symptom and mental domain by ROC analysis (P < 0.05). ROC curves also showed that the AUC of exacerbation history combined with CCQ functional domain was larger than that of exacerbation history alone (P < 0.0001). Additionally, the predictive value of multivariable models that contains CCQ functional domain was significantly better than the corresponding model without CCQ functional domain (P < 0.05). The total CCQ and its functional and symptom domain were independent risk factors of 3-year hospitalized exacerbations. The prognostic capacity of the functional domain was similar to that of total CCQ, and was significantly stronger than the symptom and mental domain. The CCQ functional domain was able to increase the predictive power of exacerbation history and other multivariable prediction models, indicating it may have an important role in the multivariable prediction tool for hospitalized exacerbation, and its combination with other clinical variables might be used as a low-cost approach for assessments of the disease severity and severe exacerbation in COPD patients in the future.

Modified local Granger causality analysis based on Peter‐Clark algorithm for multivariate time series prediction on IoT data

AbstractClimate data collected through Internet of Things (IoT) devices often contain high‐dimensional, nonlinear, and auto‐correlated characteristics, and general causality analysis methods obtain quantitative causality analysis results between variables based on conditional independence tests or Granger causality, and so forth. However, it is difficult to capture dynamic properties between variables of temporal distribution, which can obtain information that cannot be obtained by the mean detection method. Therefore, this paper proposed a new causality analysis method based on Peter‐Clark (PC) algorithm and modified local Granger causality (MLGC) analysis method, called PC‐MLGC, to reveal the causal relationships between variables and explore the dynamic properties on temporal distribution. First, the PC algorithm is applied to compute the relevant variables of each variable. Then, the results obtained in the previous stage are fed into the modified local Granger causality analysis model to explore causalities between variables. Finally, combined with the quantitative causality analysis results, the dynamic characteristic curves between variables can be obtained, and the accuracy of the causal relationship between variables can be further verified. The effectiveness of the proposed method is further demonstrated by comparing it with standard Granger causality analysis and a two‐stage causal network learning method on one benchmark dataset and two real‐world datasets.

Predictors of Durable Remission After Successful Surgery for Cushing Disease: Results From the Multicenter RAPID Registry.

Cushing disease (CD) affects mortality and quality of life along with limited long-term remission, underscoring the need to better identify recurrence risk. The identification of surgical or imaging predictors for CD remission after transsphenoidal surgery has yielded some inconsistent results and has been limited by single-center, single-surgeon, or meta-analyses studies. We sought to evaluate the multicenter Registry of Adenomas of the Pituitary and Related Disorders (RAPID) database of academic US pituitary centers to assess whether robust nonhormonal recurrence predictors could be elucidated. Patients with treated CD from 2011 to 2023 were included. The perioperative and long-term characteristics of CD patients with and without recurrence were assessed using univariable and multivariable analyses. Of 383 patients with CD from 26 surgeons achieving postoperative remission, 288 (75.2%) maintained remission at last follow-up while 95 (24.8%) showed recurrence (median time to recurrence 9.99 ± 1.34 years). Patients with recurrence required longer postoperative hospital stays (5 ± 3 vs 4 ± 2 days, P = .002), had larger average tumor volumes (1.76 ± 2.53 cm 3 vs 0.49 ± 1.17 cm 3 , P = .0001), and more often previously failed prior treatment (31.1% vs 14.9%, P = .001) mostly being prior surgery. Multivariable hazard prediction models for tumor recurrence found younger age (odds ratio [OR] = 0.95, P = .002) and Knosp grade of 0 (OR = 0.09, reference Knosp grade 4, P = .03) to be protective against recurrence. Comparison of Knosp grade 0 to 2 vs 3 to 4 showed that lower grades had reduced risk of recurrence (OR = 0.27, P = .04). Other factors such as length of stay, surgeon experience, prior tumor treatment, and Knosp grades 1, 2, or 3 failed to reach levels of statistical significance in multivariable analysis. This multicenter study centers suggests that the strongest predictors of recurrence include tumor size/invasion and age. This insight can help with patient counseling and prognostication. Long-term follow-up is necessary for patients, and early treatment of small tumors may improve outcomes.

A novel dynamic structural adaptive multivariable grey model and its application in China's solar energy generation forecasting

Accurately predicting China's solar energy generation is crucial for energy planning and environmental protection. However, the nonlinearity and complexity of the data pose challenges to prediction. This article proposes a dynamic structural adaptive multivariate grey model to overcome these challenges. The model constructs a dynamic nonlinear correction function with an uncertain structure that can adaptively adjust to the data characteristics. Additionally, the Gaussian function is utilized to derive the time response function, overcoming the mechanism defects present in traditional multivariate grey models. The whale optimization algorithm is applied to optimize the background value of the model, improving accuracy. The model is compatible with 14 classical models via parameter changes and has strong compatibility. The results show that the fitting error and test error of the model are 0.23 % and 1.95 % respectively, which is significantly better than the other 11 methods, indicative of strong robustness and accuracy. The forecast indicates that China's solar energy generation will experience a consistent upward trajectory from 2023 to 2030, with an average annual growth rate of 12.11 %. This study significantly contributes to the advancement of the theory of multivariate grey dynamic prediction and provides essential data support for strategic planning within China's solar power industry.

Multivariate Analysis and Prediction Model Construction for Dermatoscope Use by Chinese Dermatologists via an Online Survey.

The use of dermatoscopes is constantly increasing globally, but to date, there are no studies on the use of dermatoscopes by Chinese dermatologists. To determine factors influencing the use of dermatoscopes in China. A web-based questionnaire was designed by the Department of Dermatology at Huashan Hospital affiliated with Fudan University and was published online via the Shanghai Wheat Color Intelligent Technology Company, China. In 2019 and 2022, 1581 and 1507 dermatologists, respectively, were recruited and completed the questionnaire online. In China, the application rate of dermatoscopy is higher in eastern provinces than in western and remote areas. The proportion of doctors from public tertiary hospitals is the highest, with females being the majority. The age range of 30-40 years has the highest proportion, the proportion of doctors with professional titles of attending physician or above is the highest, and the proportion of doctors with a bachelor's degree or above is the highest. By improving the education and professional standards of doctors, providing more training opportunities, simplifying access, and promoting dermatoscopy in grassroots hospitals, we can increase the confidence of dermatologists in the use of dermatoscopy.

Advanced hybrid empirical mode decomposition, convolutional neural network and long short-term memory neural network approach for predicting grain pile humidity based on meteorological inputs

Grain pile humidity prediction is beneficial to ensure food security, and establishing an effective humidity prediction model is of great significance to the field of grain storage. By taking meteorological and grain temperature data as inputs, we propose a prediction model that combines Empirical Mode Decomposition (EMD), Convolutional Neural Network (CNN), and Long Short-Term Memory Network (LSTM). The model was verified in experimental data of three different storage layers of grain piles. The prediction results show that the proposed EMD-CNN-LSTM model has better prediction accuracy than the other three comparison models: CNN-LSTM, CNN and LSTM. From the average results of the entire granary, the MAE, RMSE, and MAPE results are 0.14, 0.18, and 0.25%, respectively, and the MAE value is 44% higher than the previous research method that does not consider meteorological factors. The MAE, RMSE, and MAPE results of the CNN-LSTM method with EMD decomposition were improved by 58%, 53% and 58% respectively compared with the method without EMD decomposition. It can be concluded that taking meteorological factors as model input and integrating EMD methods can improve prediction accuracy. The constructed prediction model shows effective prediction results in different storage layers of grain pile, which provides new insights for ensuring food security and also provides valuable references for multivariate time series prediction in other fields.

Stacked generalization ensemble learning strategy for multivariate prediction of delamination and maximum thrust force in composite drilling

The complexity of drilling carbon fiber reinforced polymers (CFRP) requires accurate predictive models. This study addresses the challenge using an ensemble machine learning (ML) approach with stacked generalization. The model captures the relationships between key input variables—such as graphene nanoplatelet (GNP) content, ultrasonic assistance, tool type, stacking sequence, and feed rate—and output parameters, specifically thrust force and delamination. A nested feature scoring (NFS) method was employed for importance analysis, revealing tooling type and feed rate as key features for minimizing delamination and reducing thrust force, respectively. The machinability results revealed that ultrasonic drilling lowered thrust force by improving chip evacuation and reducing fiber breakage. HSS tools with cobalt content, alongside symmetrical stacking sequence, helped to further minimize both thrust force and delamination. However, the inclusion of GNPs led to an increase in thrust force and delamination, attributed to the increased strength of the CFRP/GNP composite. The process involved meticulous training, resulting in four optimal-fit models serving as inputs for the stacked meta-model. Iterative enhancements fortified the ensemble robustness, with fine-tuning of hyperparameters through Bayesian optimization. The ensemble superiority over individual models manifested in a remarkable reduction of mean absolute error (MAE) and root mean squared error (RMSE) by up to 97% and 124% for delamination, and 205% and 154% for thrust force, compared to the best base learner. Visual and statistical assessments effectively illuminated the intricate interactions between variables in the drilling process. The methodology resulted in a highly adaptable predictive model with applications across diverse manufacturing contexts.

Risk Prediction Models for Adverse Drug Reactions and Adverse Drug Events in Older Adults – A Systematic Review and Meta-analysis

Abstract Background Adverse drug reactions (ADRs) and adverse drug events (ADEs) are common and result in significant morbidity, mortality and healthcare costs. As the world’s population ages, the consumption of medications and healthcare will increase. Models predicting ADR risks in older adults have previously been found to lack reliability and validity. The aim of this systematic review and meta-analysis is to provide an updated, comprehensive quality assessment and analysis of ADR-risk prediction tools in older adults. Methods Standard computerised-databases and citations were searched (2012 to 2023) to identify relevant peer-reviewed studies. Studies which developed and/or validated an ADR/ADE prediction model for use in older adults were included. Four studies from a previous systematic review were also included. The TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) guidelines were used to evaluate each of the included studies. Random effects models were used to derive a pooled discrimination estimate (area under the receiver operation curve; AUROC) for model development studies and model validation studies. The prediction model risk of bias assessment tool (PROBAST) was applied to evaluate bias. Results Five out of 11,481 titles, plus four previous studies, met all inclusion criteria and underwent TRIPOD evaluation. Meta-analysis resulted in a favourable pooled AUROC 0.75 (95% CI 0.57, 0.87; I-squared= 96.65%) for model development studies and 0.70 (95% CI 0.57, 0.80; I-squared= 90.53%) for model validation studies but there was substantial heterogeneity. Studies had poor adherence (range 32-50%; median 44%; IQR 11.5%) to TRIPOD guidelines. The overall risk of bias was high for all included studies. Conclusion Seven risk prediction models were identified, all exhibiting poor adherence to TRIPOD guidelines which may question the investigational rigor of the studies and their usability. This underscores the need for the development of a validated, robust and reliable tool worthy of implementation and testing in a real-world setting.

Multivariate Scalar on Multidimensional Distribution Regression With Application to Modeling the Association Between Physical Activity and Cognitive Functions

ABSTRACTWe develop a new method for multivariate scalar on multidimensional distribution regression. Traditional approaches typically analyze isolated univariate scalar outcomes or consider unidimensional distributional representations as predictors. However, these approaches are suboptimal because (i) they fail to utilize the dependence between the distributional predictors and (ii) neglect the correlation structure of the response. To overcome these limitations, we propose a multivariate distributional analysis framework that harnesses the power of multivariate density functions and multitask learning. We develop a computationally efficient semiparametric estimation method for modeling the effect of the latent joint density on the multivariate response of interest. Additionally, we introduce a new conformal prediction algorithm for quantifying the uncertainty of our multivariate predictions based on subject characteristics and individualized distributional predictors, providing valuable insights into the conditional distribution of the response. We validate the effectiveness of our proposed method through comprehensive numerical simulations, clearly demonstrating its superior performance compared to traditional methods. The application of the proposed method is demonstrated on triaxial accelerometer data from the National Health and Nutrition Examination Survey 2011–2014 for modeling the association between cognitive scores across various domains and distributional representation of physical activity among the older adult population. Our results highlight the advantages of the proposed approach, emphasizing the significance of incorporating multidimensional distributional information in the triaxial accelerometer data.