Dynamic Prediction Model Research Articles

Transfer Dynamic Latent Variable Modeling for Quality Prediction of Multimode Processes.

Quality prediction is beneficial to intelligent inspection, advanced process control, operation optimization, and product quality improvements of complex industrial processes. Most of the existing work obeys the assumption that training samples and testing samples follow similar data distributions. The assumption is, however, not true for practical multimode processes with dynamics. In practice, traditional approaches mostly establish a prediction model using the samples from the principal operating mode (POM) with abundant samples. The model is inapplicable to other modes with a few samples. In view of this, this article will propose a novel dynamic latent variable (DLV)-based transfer learning approach, called transfer DLV regression (TDLVR), for quality prediction of multimode processes with dynamics. The proposed TDLVR can not only derive the dynamics between process variables and quality variables in the POM but also extract the co-dynamic variations among process variables between the POM and the new mode. This can effectively overcome data marginal distribution discrepancy and enrich the information of the new mode. To make full use of the available labeled samples from the new mode, an error compensation mechanism is incorporated into the established TDLVR, termed compensated TDLVR (CTDLVR), to adapt to the conditional distribution discrepancy. Empirical studies show the efficacy of the proposed TDLVR and CTDLVR methods in several case studies, including numerical simulation examples and two real-industrial process examples.

Supply chain network optimization for emergency materials considering demand satisfaction

AbstractIn order to effectively guarantee the smooth supply of emergency materials under public health emergencies, first, we predict the number of daily infected people based on the modified SEIRD infectious disease dynamics model and accordingly construct the dynamic demand prediction model of emergency materials so as to catch the change rule of the demand for emergency materials with greater precision. Subsequently, under the premise of demand satisfaction, we optimize the supply chain network of emergency materials and embed two insurance strategies in the supply system in advance: multi‐source purchase and buffer production capacity, to enhance the stability of supply chain by sacrificing part of the cost. Next, a bi‐objective linear programming model based on minimizing system cost and maximizing demand satisfaction is proposed with the intention of realizing the design of emergency materials supply chain network, which balances the risk and cost. Finally, through the analysis of the example, we find that compared with the traditional supply chain network of emergency materials, the total satisfaction rate of the supply chain network optimized by balancing cost and risk is increased by 11.70%, while the cost is increased by only 2.04%, which fully verifies the validity of the optimized design of supply chain network of emergency materials.

Dynamic Simulation of the Leaf Mass per Area (LMA) in Multilayer Crowns of Young Larix principis-rupprechtii.

Leaf mass per area (LMA) is a key structural parameter that reflects the functional traits of leaves and plays a vital role in simulating the material and energy cycles of plant ecosystems. In this study, vertical whorl-by-whorl sampling of LMA was conducted in a young Larix principis-rupprechtii plantation during the growing season at the Saihanba Forest Farm. The vertical and seasonal variations in LMA were analysed. Subsequently, a predictive model of LMA was constructed. The results revealed that the LMA varied significantly between different crown whorls and growing periods. In the vertical direction of the crown, the LMA decreased with increasing crown depth, but the range of LMA values from the tree top to the bottom was, on average, 30.4 g/m2, which was approximately 2.5 times greater in the fully expanded phase than in the early leaf-expanding phase. During different growing periods, the LMA exhibited an allometric growth trend that increased during the leaf-expanding phase and then tended to stabilize. However, the range of LMA values throughout the growing period was, on average, 40.4 g/m2. Among the univariate models, the leaf dry matter content (LDMC) performed well (adjusted determination coefficient (Ra2) = 0.45, root mean square error (RMSE) = 13.48 g/m2) in estimating the LMA. The correlation between LMA and LDMC significantly differed at different growth stages and at different vertical crown whorls. The dynamic predictive model of LMA constructed with the relative depth in the crown (RDINC) and date of the year (DOY) as independent variables was reliable in both the assessments (Ra2 = 0.68, RMSE = 10.25 g/m2) and the validation (absolute mean error (MAE) = 8.05 g/m2, fit index (FI) = 0.682). Dynamic simulations of crown LMA provide a basis for elucidating the mechanism of crown development and laying the foundation for the construction of an ecological process model.

Predicting lung function decline in cystic fibrosis: the impact of initiating ivacaftor therapy

BackgroundModulator therapies that seek to correct the underlying defect in cystic fibrosis (CF) have revolutionized the clinical landscape. Given the heterogeneous nature of lung disease progression in the post-modulator era, there is a need to develop prediction models that are robust to modulator uptake.MethodsWe conducted a retrospective longitudinal cohort study of the CF Foundation Patient Registry (N = 867 patients carrying the G551D mutation who were treated with ivacaftor from 2003 to 2018). The primary outcome was lung function (percent predicted forced expiratory volume in 1 s or FEV1pp). To characterize the association between ivacaftor initiation and lung function, we developed a dynamic prediction model through covariate selection of demographic and clinical characteristics. The ability of the selected model to predict a decline in lung function, clinically known as an FEV1-indicated exacerbation signal (FIES), was evaluated both at the population level and individual level.ResultsBased on the final model, the estimated improvement in FEV1pp after ivacaftor initiation was 4.89% predicted (95% confidence interval [CI]: 3.90 to 5.89). The rate of decline was reduced with ivacaftor initiation by 0.14% predicted/year (95% CI: 0.01 to 0.27). More frequent outpatient visits prior to study entry and being male corresponded to a higher overall FEV1pp. Pancreatic insufficiency, older age at study entry, a history of more frequent pulmonary exacerbations, lung infections, CF-related diabetes, and use of Medicaid insurance corresponded to lower FEV1pp. The model had excellent predictive accuracy for FIES events with an area under the receiver operating characteristic curve of 0.83 (95% CI: 0.83 to 0.84) for the independent testing cohort and 0.90 (95% CI: 0.89 to 0.90) for 6-month forecasting with the masked cohort. The root-mean-square errors of the FEV1pp predictions for these cohorts were 7.31% and 6.78% predicted, respectively, with standard deviations of 0.29 and 0.20. The predictive accuracy was robust across different covariate specifications.ConclusionsThe methods and applications of dynamic prediction models developed using data prior to modulator uptake have the potential to inform post-modulator projections of lung function and enhance clinical surveillance in the new era of CF care.

Proactive Identification of Patients with Diabetes at Risk of Uncontrolled Outcomes during a Diabetes Management Program: Conceptualization and Development Study Using Machine Learning.

The growth in the capabilities of telehealth have made it possible to identify individuals with a higher risk of uncontrolled diabetes and provide them with targeted support and resources to help them manage their condition. Thus, predictive modeling has emerged as a valuable tool for the advancement of diabetes management. This study aimed to conceptualize and develop a novel machine learning (ML) approach to proactively identify participants enrolled in a remote diabetes monitoring program (RDMP) who were at risk of uncontrolled diabetes at 12 months in the program. Registry data from the Livongo for Diabetes RDMP were used to design separate dynamic predictive ML models to predict participant outcomes at each monthly checkpoint of the participants' program journey (month-n models) from the first day of onboarding (month-0 model) up to the 11th month (month-11 model). A participant's program journey began upon onboarding into the RDMP and monitoring their own blood glucose (BG) levels through the RDMP-provided BG meter. Each participant passed through 12 predicative models through their first year enrolled in the RDMP. Four categories of participant attributes (ie, survey data, BG data, medication fills, and health signals) were used for feature construction. The models were trained using the light gradient boosting machine and underwent hyperparameter tuning. The performance of the models was evaluated using standard metrics, including precision, recall, specificity, the area under the curve, the F1-score, and accuracy. The ML models exhibited strong performance, accurately identifying observable at-risk participants, with recall ranging from 70% to 94% and precision from 40% to 88% across the 12-month program journey. Unobservable at-risk participants also showed promising performance, with recall ranging from 61% to 82% and precision from 42% to 61%. Overall, model performance improved as participants progressed through their program journey, demonstrating the importance of engagement data in predicting long-term clinical outcomes. This study explored the Livongo for Diabetes RDMP participants' temporal and static attributes, identification of diabetes management patterns and characteristics, and their relationship to predict diabetes management outcomes. Proactive targeting ML models accurately identified participants at risk of uncontrolled diabetes with a high level of precision that was generalizable through future years within the RDMP. The ability to identify participants who are at risk at various time points throughout the program journey allows for personalized interventions to improve outcomes. This approach offers significant advancements in the feasibility of large-scale implementation in remote monitoring programs and can help prevent uncontrolled glycemic levels and diabetes-related complications. Future research should include the impact of significant changes that can affect a participant's diabetes management.

A novel dynamic grey multivariate prediction model for multiple cumulative time-delay shock effects and its application in energy emission forecasting

Energy emission systems are often influenced by external information and policy shock. However, conventional grey models ignore the existence of the impact of multiple shock events and cumulative time-delay effects, which are crucial for achieving accurate forecasts. From a dynamic perspective, we aim to establish a multivariable grey prediction model considering the impact of multiple shock events, namely MSGTDM(1,N) model. Specifically, to accurately describe the impact of different types of external shocks on the system, three kinds of nonlinear dynamic shock functions are designed, including growth dynamic shock function, decline dynamic shock function, and slope dynamic shock function. Based on this, multi-dimensional instantaneous shock utility terms and cumulative time-delay utility terms are constructed to accurately measure shock effects. And Whale Optimization Algorithm is employed to determine the optimal parameters of the shock functions through comprehensive comparative analysis. The findings affirm the MSGTDM(1,N) model's higher predictive accuracy and validity. Consequently, the forecast results of China's carbon emissions from coal and natural gas consumption in 2022–2025 provide a reliable basis for adjusting the energy structure and implementing the “dual-carbon” policy effectively.

A dynamic prediction model of landslide displacement based on VMD–SSO–LSTM approach

Addressing the limitations of existing landslide displacement prediction models in capturing the dynamic characteristics of data changes, this study introduces a novel dynamic displacement prediction model for landslides. The proposed method combines Variational Mode Decomposition (VMD) with Sparrow Search Optimization (SSO) and Long Short-Term Memory (LSTM) techniques to formulate a comprehensive VMD–SSO–LSTM model. Through the application of VMD, the method dissects cumulative displacement and rainfall data, thereby extracting distinct components such as trend, periodicity, and fluctuation components for displacement, as well as low-frequency and high-frequency components for rainfall. Furthermore, leveraging Gray Correlational Analysis, the interrelationships between the periodic component of displacement and the low-frequency component of rainfall, as well as the fluctuation component of displacement and the high-frequency component of rainfall, are established. Building upon this foundation, the SSO–LSTM model dynamically predicts the interrelated displacement components, synthesizing the predicted values of each component to generate real-time dynamic forecasts. Simulation results underscore the effectiveness of the proposed VMD–SSO–LSTM model, indicating root-mean-square error (RMSE) and mean absolute percentage error (MAPE) values of 1.2329 mm and 0.1624%, respectively, along with a goodness of fit (R2) of 0.9969. In comparison to both back propagation (BP) prediction model and LSTM prediction model, the VMD–SSO–LSTM model exhibits heightened predictive accuracy.

A web-based dynamic predictive model for postoperative nausea and vomiting in patient receiving gynecological laparoscopic surgery.

The aim of this study was to develop a web-based dynamic prediction model for postoperative nausea and vomiting (PONV) in patients undergoing gynecologic laparoscopic surgery. The patients (N = 647) undergoing gynecologic laparoscopic surgery were included in this observational study. The candidate risk-factors related to PONV were included through literature search. Lasso regression was utilized to screen candidate risk-factors, and the variables with statistical significance were selected in multivariable logistic model building. The web-based dynamic Nomogram was used for model exhibition. Accuracy and validity of the experimental model (EM) were evaluated by generating receiver operating characteristic (ROC) curves and calibration curves. Hosmer-Lemeshow test was used to evaluate the goodness of fit of the model. Decision curve analysis (DCA) was used to evaluate the clinical practicability of the risk prediction model. Ultimately, a total of five predictors including patient-controlled analgesia (odds ratio [OR], 4.78; 95% confidence interval [CI], 1.98-12.44), motion sickness (OR, 4.80; 95% CI, 2.71-8.65), variation of blood pressure (OR, 4.30; 95% CI, 2.41-7.91), pregnancy vomiting history (OR, 2.21; 95% CI, 1.44-3.43), and pain response (OR, 1.64; 95% CI, 1.48-1.83) were selected in model building. Assessment of the model indicates the discriminating power of EM was adequate (ROC-areas under the curve, 93.0%; 95% CI, 90.7%-95.3%). EM showed better accuracy and goodness of fit based on the results of the calibration curve. The DCA curve of EM showed favorable clinical benefits. This dynamic prediction model can determine the PONV risk in patients undergoing gynecologic laparoscopic surgery.

Development and internal validation of a dynamic fall risk prediction and monitoring tool in aged care using routinely collected electronic health data: a landmarking approach.

Falls pose a significant challenge in residential aged care facilities (RACFs). Existing falls prediction tools perform poorly and fail to capture evolving risk factors. We aimed to develop and internally validate dynamic fall risk prediction models and create point-based scoring systems for residents with and without dementia. A longitudinal cohort study using electronic data from 27 RACFs in Sydney, Australia. The study included 5492 permanent residents, with a 70%-30% split for training and validation. The outcome measure was the incidence of falls. We tracked residents for 60 months, using monthly landmarks with 1-month prediction windows. We employed landmarking dynamic prediction for model development, a time-dependent area under receiver operating characteristics curve (AUROCC) for model evaluations, and a regression coefficient approach to create point-based scoring systems. The model identified 15 independent predictors of falls in dementia and 12 in nondementia cohorts. Falls history was the key predictor of subsequent falls in both dementia (HR 4.75, 95% CI, 4.45-5.06) and nondementia cohorts (HR 4.20, 95% CI, 3.87-4.57). The AUROCC across landmarks ranged from 0.67 to 0.87 for dementia and from 0.66 to 0.86 for nondementia cohorts but generally remained between 0.75 and 0.85 in both cohorts. The total point risk score ranged from -2 to 57 for dementia and 0 to 52 for nondementia cohorts. Our novel risk prediction models and scoring systems provide timely person-centered information for continuous monitoring of fall risk in RACFs. Embedding these tools within electronic health records could facilitate the implementation of targeted proactive interventions to prevent falls.

Two‐layer dynamic economic nonlinear model predictive control for a Lithium‐ion battery charge process with random disturbances

AbstractThe charging process of lithium‐ion batteries is necessary for normal operation, and improper lithium‐ion battery charging strategy can cause side reactions, significant temperature rise, performance degradation, and safety concerns. This paper proposes a two‐layer dynamic economic nonlinear model predictive control economic model predictive control (EMPC) method for lithium‐ion battery charge management in which the non‐Gaussian random noise of voltage and current signal are taken into account. In the two‐layer EMPC, the upper layer finds the optimal reference trajectory, and the power in the future prediction horizon is chosen as an economic indicator to optimize the upper layer. The lower layer tracks the optimal trajectory, and model predictive control based on generalized correntropy is used. The economic cost of the system and the dynamic changes of the process are considered to greatly reduce the computational complexity and realize rapid battery charge management. Finally, the simulation results show that the tracking error of the two‐layer EMPC method based on generalized correntropy is stable around 0, while the tracking error of the two‐layer EMPC method based on MSE is stable around 0.003. The average control action times of the proposed two‐layer EMPC and single‐layer EMPC are 0.0052 and 0.0267 s, respectively. It is verified that the proposed method performs better for the lithium‐ion battery charging process with random disturbances.

Acute Pancreatitis in Pregnancy: A Propensity Score Matching Analysis and Dynamic Nomogram for Risk Assessment.

Acute pancreatitis is easily confused with abdominal pain symptoms, and it could lead to serious complications for pregnant women and fetus, the mortality was as high as 3.3% and 11.6-18.7%, respectively. However, there is still lack of sensitive laboratory markers for early diagnosis of APIP and authoritative guidelines to guide treatment. The purpose of this study was to explore the risk factors of acute pancreatitis in pregnancy, establish, and evaluate the dynamic prediction model of risk factors in acute pancreatitis in pregnancy patients. Clinical data of APIP patients and non-pregnant acute pancreases patients who underwent regular antenatal check-ups during the same period were collected. The dataset after propensity matching was randomly divided into training set and verification set at a ratio of 7:3. The model was constructed using Logistic regression, least absolute shrinkage and selection operator regression, R language and other methods. The training set model was used to construct the diagnostic nomogram model and the validation set was used to validate the model. Finally, the accuracy and clinical practicability of the model were evaluated. A total of 111 APIP were included. In all APIP patients, hyperlipidemic pancreatitis was the most important reason. The levels of serum amylase, creatinine, albumin, triglyceride, high-density lipoprotein cholesterol, and apolipoprotein A1 were significantly different between the two groups. The propensity matching method was used to match pregnant pancreatitis patients and pregnant non-pancreatic patients 1:1 according to age and gestational age, and the matching tolerance was 0.02. The multivariate logistic regression analysis of training set showed that diabetes, triglyceride, Body Mass Index, white blood cell, and C-reactive protein were identified and entered the dynamic nomogram. The area under the ROC curve of the training set was 0.942 and in validation set was 0.842. The calibration curve showed good predictive in training set, and the calibration performance in the validation set was acceptable. The calibration curve showed the consistency between the nomogram model and the actual probability. The dynamic nomogram model we constructed to predict the risk factors of acute pancreatitis in pregnancy has high accuracy, discrimination, and clinical practicability.

Dynamic Fractional-Order Grey Prediction Model with GWO and MLP for Forecasting Overseas Talent Mobility in China

International students play a crucial role in China’s talent development strategy. Thus, predicting overseas talent mobility is essential for formulating scientifically reasonable talent introduction policies, optimizing talent cultivation systems, and fostering international talent cooperation. In this study, we proposed a novel fractional-order grey model based on the Multi-Layer Perceptron (MLP) and Grey Wolf Optimizer (GWO) algorithm to forecast the movement of overseas talent, namely MGDFGM(1,1). Compared to the traditional grey model FGM(1,1), which utilizes the same fractional order at all time points, the proposed MGDFGM(1,1) model dynamically adjusts the fractional-order values based on the time point. This dynamic adjustment enables our model to better capture the changing trends in the data, thereby enhancing the model’s fitting capability. To validate the effectiveness of the MGDFGM(1,1) model, we primarily utilize Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) as the evaluation criteria for the prediction accuracy, as well as standard deviation (STD) as an indicator of the model stability. Furthermore, we perform experimental analysis to evaluate the predictive performance of the MGDFGM(1,1) model in comparison to NAÏVE, ARIMA, GM(1,1), FGM(1,1), LSSVR, MLP, and LSTM. The research findings demonstrate that the MGDFGM(1,1) model achieves a remarkably high level of prediction accuracy and stability for forecasting overseas talent mobility in China. The implications of this study offer valuable insights and assistance to government departments involved in overseas talent management.

Dynamic Measurement and Prediction of Sulfur Hexafluoride Gas Weight Based on Non-Ideal Gas State Equation and Improved Random Forest

Real-time knowledge of sulfur hexafluoride (SF6) gas weight can effectively manage and maintain the power equipment using SF6. Firstly, the gas pressure and temperature sensors are utilized to measure SF6 gas pressure and temperature dynamically. Secondly, the measured gas pressure and temperature values are converted to the SF6 gas weight using equations for calculating non-ideal gas states. Finally, the dynamic prediction model of the SF6 gas weight is developed by combining the improved chimp optimization algorithm and random forest algorithm. The dynamic prediction model is trained using measured pressure and temperature data as inputs and the calculated SF6 gas weight as outputs. The trained prediction model enables the dynamic prediction of SF6 gas weight. The experimental results show that the mean absolute error, root mean square error, and mean absolute percentage error of the proposed prediction model reaches 51.54 kg, 103.59 kg, and 11.62%.

A Dynamic Prediction Framework for Urban Public Space Vitality: From Hypothesis to Algorithm and Verification

Predicting and assessing the vitality of public urban spaces is crucial for effective urban design, aiming to prevent issues such as “ghost streets” and minimize resource wastage. However, existing assessment methods often lack temporal dynamics or heavily rely on historical big data, limiting their ability to accurately predict outcomes for unbuilt projects. To address these challenges, this study integrates previous methodologies with observations of crowd characteristics in public spaces. It introduces the crowd-frequency hypothesis and develops an algorithm to establish a time-dimensional urban vitality dynamic prediction model. Through a case study of the Rundle Mall neighborhood in Adelaide, Australia, the effectiveness of the prediction model was validated using on-site observation sampling and comparative verification. The prediction model framework allows for the determination of urban vitality within specific time ranges by directly inputting basic information, providing valuable support to urban planners and government officials during the design and decision-making processes. It offers a cost-effective approach to achieve sustainable urban vitality construction. Furthermore, machine learning techniques, specifically the decision tree model, were applied to case data to develop a set of preliminary algorithm tools, which enable output of reference urban vitality levels (high-medium-low).

Reliable Cost Prediction and Control for Intelligent Manufacture: A Key Performance Indicator Perspective

Article Reliable Cost Prediction and Control for Intelligent Manufacture: A Key Performance Indicator Perspective Hang Geng 1,*, Alireza Mousavi 2, Nikolaos Grigorios Markatos 2, Kai Chen 1, and Xuan Gou 1 1 School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China 2 Department of Computer Science, Brunel University London, Uxbridge, Middlesex, UB8 3PH, United Kingdom * Correspondence: Hang.Geng@uestc.edu.cn Received: 10 July 2023 Accepted: 11 October 2023 Published: 26 March 2024 Abstract: Intelligent manufacturing is facing significant challenges in adapting to the ever-changing equipment, instrumentation, process and economics. Such a trend together with the pressure to reliably control and contain production costs means that frequent adjusting decisions are required to adapt to incessant volatility imposed on manufacturing systems. Under this circumstance, cost-effective and quality-guaranteed manufacturing strategies would be the most logical route to reducing production costs. In this paper, a novel dynamical cost prediction and control (CPC) model is proposed to support collective decision-making in intelligent manufacturing, where the model output is the real-time prediction of possible manufacturing costs, while the inputs are generic manufacturing key performance indicators covering inventory, product quality, production efficiency, resource utilisation and environmental impact. This proposed CPC model distinguishes itself from existing ones for its capability to translate manufacturing data (at both the physical level and operation management level) into financial metrics that contribute to forming a common language between engineering, financial and administrative departments of an enterprise. The case study about the assembly line of optoelectronic devices demonstrates that, although different enterprise departments have different priorities, our CPC model helps them to achieve certain consensus on intended production that finally creates satisfactory profitability for the company at controlled manufacturing costs.

Impact factors and peaking simulation of carbon emissions in the building sector in Shandong Province

The building sector has the fastest growth rate of carbon emissions and the greatest potential for carbon reduction. The building sector in Shandong Province ranks first in China in terms of total carbon emissions and growth rate. The study of carbon emissions in the building sector in Shandong Province is an essential reference for Chinese building sector to achieve the “carbon peaking and carbon neutrality” target. This study uses historical data on carbon emissions in the building sector in Shandong Province, based on which the influencing factors and carbon reduction potential are analyzed. A dynamic and comprehensive carbon emission prediction model for buildings was developed to predict the trajectory of carbon emissions in the building sector in Shandong Province from 2020 to 2050 with the corresponding probability distribution. Results show that the main factors driving the increase in carbon emissions in the building sector in Shandong Province are the level of affluence of the residents and the level of building services. The building area demand index and building energy efficiency are the most important factors inhibiting the increase in carbon emissions in the building sector in Shandong Province during the same period. Under the baseline scenario, carbon emissions in the building sector in Shandong Province will peak at 243.573 (±46.022) million tonnes CO2 in 2035 (±3) years. The floor area demand index and level of affluence of the residents are the most critical factors affecting the uncertainty of future of carbon emissions in the building sector in Shandong Province.

Abstract 6619: Dynamic prediction with repeated mammogram images improves 5-year breast cancer risk performance

Abstract Background: Increasing use of digital screening mammography opens potential to add individualized dynamic measures of breast image history to predict future breast cancer risk. Current prediction models use a single mammogram measure, usually breast density, at one point in time, to add to prediction models based on established risk factors. Our previous work has shown that change in each breast over time is related to risk of diagnosis with breast cancer. Methods: We apply a dynamic prediction model that includes full field digital mammogram in each breast to predict future 5-year risk. Data from Joanne Knight Breast Health Cohort (JKBHC) serve as training population with 290 pathology-confirmed breast cancer cases and 658 controls sampled with a total number of 8,712 craniocaudal (CC) view mammograms for analysis. Nurses’ Health Study 2 (NHS2) serves as the external validation with 188 cases and 462 controls sampled with a total number of 3,712 CC-views. The dynamic risk prediction model aims to estimate the probability of future breast cancer risk conditional on the past history of the longitudinal mammograms. We train the model on the JKBH cohort, retaining 20% for testing and 15% for validation. We report a time dependent AUC when evaluating model performance. Results: The mean age of women was 56.6 years at entry to cohort; 81.3% were white. Adding trajectory of mammogram images in each breast to a model improved performance. The mean number of mammograms conditional on 5 years of observation in JKBHC was 3.36 (1.49) and in NHS2 2.36 (1.53). We obtained a 5-year dynamic AUC of 0.81 (0.05) in the JKBHC internal validation and 0.77 (0.06) in the NHS2 external validation. Model performance was comparable in pre- and postmenopausal women. Because breasts with high mammographic density hinder detection of tumors we stratify analyses by dense and non-dense at baseline. Among women with dense breasts at baseline (BIRADS C and D), we observe a dynamic 5-year risk AUC = 0.77. In the subgroup with less dense breasts (BIRADS A and B) we observe AUC = 0.74. Conclusions: With this approach we can accommodate the irregularly observed time-varying images and project only partial information prior to doing the prediction. Adding trajectory of mammogram history in each breast improves 5-year breast cancer risk prediction beyond models used in clinic today. Citation Format: Graham A. Colditz, Rulla M. Tamimi, Bernard Rosner, Shu Jiang. Dynamic prediction with repeated mammogram images improves 5-year breast cancer risk performance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6619.

Research on adaptive prediction model of rate of penetration under dynamic formation conditions

Rate of Penetration (ROP) prediction is a crucial for optimizing drilling processes and enhancing drilling efficiency. However, most existing ROP prediction models are static, relying on fixed neighboring well datasets and incapable of adapting to real-time changes in formation conditions during drilling. The dynamic and uncertain nature of geological conditions and formation heterogeneity further limits the extension and application of current ROP prediction models to target wells. To address these challenges, this study introduces an Adaptive Random Forest (ARF) algorithm to establish an ROP adaptive prediction model under dynamic formation conditions during real-time drilling operations. The performance of the ARF model is compared with the Moving-Window Random Forest (MW-RF) ROP dynamic prediction model, using data from the W1 well in the Sichuan basin as an example. Experimental results demonstrate that the ARF model outperforms the MW-RF model, exhibiting smaller overall error and reduced error fluctuation during formation condition changes. The ARF model's dynamic adaptability allows it to respond to the evolving trends of the ROP prediction model, rapidly adjusting to changes in formation conditions based on real-time drilling data flow. In contrast, the MW-RF model exhibits slower adaptability and delays in updating, reacting passively to window movements. The proposed ARF model overcomes limitations associated with static ROP prediction models, making it applicable to practical drilling scenarios and providing a theoretical foundation for real-time drilling parameter optimization and intelligent drilling strategies.

DGAMDA: Predicting miRNA-disease association based on dynamic graph attention network.

MiRNA (microRNA)-disease association prediction has essential applications for early disease screening. The process of traditional biological experimental validation is both time-consuming and expensive. However, as artificial intelligence technology continues to advance, computational methods have become efficient tools for predicting miRNA-disease associations. These methods often rely on the combination of multiple sources of association data and require improved feature mining. This study proposes a dynamic graph attention-based association prediction model, DGAMDA, which combines feature mapping and dynamic graph attention mechanisms through feature mining on a single miRNA-disease association network. DGAMDA effectively solves the problems of feature heterogeneity and inadequate feature mining by previous static graph attention mechanisms and achieves high-precision feature mining and association scoring prediction. We conducted a five-fold cross-validation experiment and obtained the mean values of Accuracy, Precision, Recall, and F1-score, which were .8986, .8869, .9115, and .8984, respectively. Our proposed model outperforms other advanced models in terms of experimental results, demonstrating its effectiveness in feature mining and association prediction based on a single association network. In addition, our model can also be used to predict miRNAs associated with unknown diseases.

GNN-LSTM-based fusion model for structural dynamic responses prediction

With the rapid growth of deep learning technology, the potential for its use in structural engineering has substantially increased in recent years. This study proposes an innovative deep-learning fusion network architecture based on the graph neural network (GNN) and long short-term memory (LSTM) network. The proposed fusion model can accurately predict nonlinear floor acceleration, velocity, and displacement responses of typical steel moment-resisting frames (SMRF) with 4 through 7 stories subjected to strong ground motions. The fusion model framework overcomes a major drawback in existing deep learning models as it can predict the dynamic responses of various structures. This has widened the range of potential applications of structural surrogate models generated through deep learning technology on the design and analysis of building structures with different geometries. Additionally, this paper presents two LSTM-optimized learning strategies, namely packing padded sequences and sequences compression strategies. These strategies improve the model’s performance significantly without modifying its architecture. Finally, this study reveals that the model’s internal graph embedding is highly correlated with certain critical structural parameters, such as the first natural period and the height of a building. This shows that the proposed fusion model is interpretable and has the ability to extract vital information that influences structural dynamic responses.