Complex Nonlinear Relationship Research Articles

A novel method for ship carbon emissions prediction under the influence of emergency events

Accurate prediction of ship emissions aids to ensure maritime sustainability but encounters challenges, such as the absence of high-precision and high-resolution databases, complex nonlinear relationships, and vulnerability to emergency events. This study addresses these issues by developing novel solutions: a novel Spatiotemporal Trajectory Search Algorithm (STSA) based on Automatic Identification System (AIS) data; a rolling structure-based Seasonal-Trend decomposition based on the Loess technique (STL); a modular deep learning model based on Structured Components, stacked-Long short-term memory, Convolutional neural networks and Comprehensive forecasting module (SCLCC). Based on these solutions, a case study using pre and post-COVID-19 AIS data demonstrates model reliability and the pandemic’s impact on ship emissions. Numerical experiments reveal that the STSA algorithm significantly outperforms the conventional identification standard in terms of accuracy of ship navigation state identification; the SCLCC model exhibits greater resistance against emergency events and excels in comprehensively capturing global information, thus yielding higher accurate prediction results. This study sheds light on the changing dynamics of maritime transport and its impacts on carbon emissions.

Machine learning-based prediction of cadmium pollution in topsoil and identification of critical driving factors in a mining area.

Mining activities have resulted in a substantial accumulation of cadmium (Cd) in agricultural soils, particularly in southern China. Long-term Cd exposure can cause plant growth inhibition and various diseases. Rapid identification of the extent of soil Cd pollution and its driving factors are essential for soil management and risk assessment. However, traditional geostatistical methods are difficult to simulate the complex nonlinear relationships between soil Cd and potential features. In this study, sequential extraction and hotspot analyses indicated that Cd accumulation increased significantly near mining sites and exhibited high mobility. The concentration of Cd was estimated using three machine learning models based on 3169 topsoil samples, seven quantitative variables (soil pH, Fe, Ca, Mn, TOC, Al/Si and ba value) and three quantitative variables (soil parent rock, terrain and soil type). The random forest model achieved marginally better performance than the other models, with an R2 of 0.78. Importance analysis revealed that soil pH and Ca and Mn contents were the most significant factors affecting Cd accumulation and migration. Conversely, due to the essence of controlling Cd migration being soil property, soil type, terrain, and soil parent materials had little impact on the spatial distribution of soil Cd under the influence of mining activities. Our results provide a better understanding of the geochemical behavior of soil Cd in mining areas, which could be helpful for environmental management departments in controlling the diffusion of Cd pollution and capturing key targets for soil remediation.

Optimal design of an inclined barrier for performance improvement of an oscillating water column device with an artificial neural network model

An oscillating water column (OWC) device with a surface-piercing inclined barrier has been studied in irregular waves. This study is to determine the optimal shape of an inclined barrier utilizing an Artificial Neural Network (ANN) model for the performance enhancement of the conventional OWC system. We employed the dual boundary element method (DBEM) as our numerical approach, which resolves two separate boundary integral equations (singular and hyper-singular) formulated at the coincident source points situated on either side of the degenerate barrier boundary. The design parameters affecting the performance of the OWC system have been optimized to maximize the power extraction using the trained ANN model. The training dataset was prepared using the DBEM, which is then used to train and validate the ANN model. Subsequently, the trained model was employed to make predictions using a broader dataset to identify the optimal combination of design parameters that maximize the conversion efficiency. The findings reveal that the well-trained ANN model can learn complex nonlinear relationships, make good predictions with high accuracy using a larger dataset of input conditions, and offer the optimal values of an inclined barrier appropriate for the wave conditions at the sea installation site.

How does working time impact perceived mental disorders? New insights into the U-shaped relationship.

Based on a large-scale nationally representative survey in China, this paper uses the exogenous impact of automation on working hours as the instrumental variable to examine working time's impact on perceived mental disorders, on the basis of dealing with endogeneity. Different from existing literature, it is found that the impact of working time on perceived mental disorders is U-shaped, rather than linear. Mental disorders firstly decrease with working hours. After working more than 48.688 h per week, further increases in working time carry notable mental health costs, leading to a positive relationship between working hours and depression. The turning point of this U-shaped relationship is almost in line with the International Labor Organization's 48 working hours/week standard, justifying it from a mental health perspective. In addition, we further exclude the possibility of more complex nonlinear relationships between working time and perceived mental disorders. Furthermore, heterogeneities are found in the effects of working hours on mental disorders across different subgroups. Males are more depressed when working overtime. Older workers have a lower tolerance for overwork stress. The turning point is smaller for the highly educated group and they are more sensitive to working longer. Those with higher socioeconomic status are less depressed after exceeding the optimal hours of work. The increase in depression among rural workers faced with overwork is not prominent. Perceived mental disorders are lower among immigrants and those with higher health status. In addition, labor protection and social security help to weaken mental disorders caused by overtime work. In conclusion, this paper demonstrates that working time has a U-shaped impact on perceived mental disorders and highlights the vulnerability of certain groups, providing a reference for setting optimal working hours from a mental health perspective.

Early Postoperative Prediction of Complications and Readmission After Colorectal Cancer Surgery Using an Artificial Neural Network.

Early predictors of postoperative complications can risk-stratify patients undergoing colorectal cancer surgery. However, conventional regression models have limited power to identify complex nonlinear relationships among a large set of variables. We developed artificial neural network models to optimize the prediction of major postoperative complications and risk of readmission in patients undergoing colorectal cancer surgery. The aim of this study was to develop an artificial neural network model to predict postoperative complications using postoperative laboratory values, and compare these models' accuracy to standard regression methods. This retrospective study included patients who underwent elective colorectal cancer resection between January 1, 2016, and July 31, 2021. Clinical data, cancer stage, and laboratory data from postoperative day 1 to 3 were collected. Models of complications and readmission risk were created using multivariable logistic regression and single-layer neural networks. National Cancer Institute-Designated Comprehensive Cancer Center. Adult colorectal cancer patients. Accuracy of predicting postoperative major complication, readmission and anastomotic leak using the area under the receiver-operating characteristic curve. Neural networks had larger areas under the curve for predicting major complications compared to regression models (neural network 0.811; regression model 0.724, p < 0.001). Neural networks also showed an advantage in predicting anastomotic leak (p = 0.036) and readmission using postoperative day 1-2 values (p = 0.014). Single-center, retrospective design limited to cancer operations. In this study, we generated a set of models for early prediction of complications after colorectal surgery. The neural network models provided greater discrimination than the models based on traditional logistic regression. These models may allow for early detection of postoperative complications as soon as postoperative day 2. See Video Abstract.

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.

Utilizing a CNN-RNN machine learning approach for forecasting time-series outlet fluid temperature monitoring by long-term operation of BHEs system

The Borehole Heat Exchanger (BHE) plays a pivotal role in enhancing heat exchange efficiency within Ground Source Heat Pump (GSHP) systems. The accurate prediction of the BHE's outlet fluid temperature is crucial for optimizing GSHP performance, energy storage, and resource conservation. However, conventional machine learning methods encounter challenges in manual feature extraction, learning complex nonlinear relationships, and adapting to real-world scenarios. To address these limitations, this research proposes a crossbreed model integrating Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) architectures to forecast long-term outlet fluid temperature in BHE systems. The model framework encompasses data preprocessing, utilizing refined data in the CNN module for temporal feature extraction, subsequently passed to the RNN module to capture sequential and temporal patterns from each dataset. Specifically, the advanced CNN-RNN architecture is designed to establish a comprehensive input-output mapping, leveraging essential input features such as inlet fluid, ambient air, and subsurface temperatures at varying depths (0, 10, and 20 m). Performance evaluation metrics, including R2, RMSE, MAE, and AARE, are employed to compare and assess prediction accuracy across various models, including LSTM, CNN, and SimpleRNN. The obtained results demonstrate the superior performance of the proposed model, achieving an RSME of 0.818, MAE of 0.642, AARE of 0.0305, and an R2 value of 98.75 %. This surpasses the performance of traditional prediction models (LSTM, CNN, and SimpleRNN) by 3.01 %, 5.80 %, and 19.52 %, respectively. Notably, the remarkably low MAE of 0.642 exhibited by a CNN-RNN model underscores its capability to outperform traditional approaches, especially when handling large datasets. These findings emphasize the significance of the developed model in facilitating efficient operation, positioning it as a valuable tool for advancing the long-term sustainability of BHE systems.

Multi-cluster nonlinear unsupervised feature selection via joint manifold learning and generalized Lasso

Due to the scarcity of data labels, unsupervised feature selection has received a lot of attention in recent years. While many unsupervised feature selection methods are capable of selecting relevant features, they often fail to comprehensively consider the impact of both local and global information of the data on feature selection, nor can they effectively handle the complex nonlinear relationships commonly found in real-world data. As a result, suboptimal feature subsets are often selected. In this paper, inspired by the Uniform Manifold Approximation and Projection (UMAP) manifold learning technique and the nonlinear sparse learning method based on Feature-Wise Kernelized Lasso, we propose a novel unsupervised feature selection method called Multi-Cluster Unsupervised Nonlinear Feature Selection based on UMAP and block HSIC Lasso (MUNFS). MUNFS greatly improves the representation of high-dimensional data during dimensionality reduction and effectively handles complex nonlinear relationships in such data. Specifically, by capturing the intrinsic topology of the data, MUNFS accurately preserves the local structure of the data while keeping as much of the global structure as possible. Furthermore, the kernel-based Hilbert–Schmidt Independence Criterion (HSIC) may measure the nonlinear dependency between the features and the target variables, while applying the l1 regularization term in feature selection to achieve sparsity. This allows for a more precise assessment of the significance of each feature. Extensive experimental results on five benchmark datasets and eight hyperspectral datasets demonstrate that the MUNFS method performs much better than several other feature selection methods.

Deep‐learning‐based downscaling of precipitation in the middle reaches of the Yellow River using residual‐based CNNs

AbstractThe middle reaches of the Yellow River (MRYR), located in northern China, are the transition zone between semi‐arid and semi‐humid climates. As one of the climate‐sensitive regions in China, MRYR has a fragile ecological environment and serious soil loss, which leads to geological disasters such as landslides, collapses, and mudslides caused by extreme precipitation. However, scarceness of high‐resolution precipitation data over MRYR limits assessment of the environmental impacts caused by climate change, especially for extreme precipitation. In this article, we design a Residual‐in‐Residual Dense Block based Network (RRDBNet) model for the statistical downscaling of precipitation in MRYR, and compare the proposed RRDBNet with a generalized linear regression model (GLM) and two popular deep‐learning‐based models. The multi‐level residuals and dense connectivity strategies introduced in RRDBNet help it to learn more abstract features and complex nonlinear relationships among climate variables to improve downscaling performance. The results show that the proposed RRDBNet has good performance in precipitation simulations, which can reproduce the spatial–temporal characteristics of high‐resolution precipitation well. RRDBNet reduces the root‐mean‐squared error (RMSE) by 19% and improves the Pearson correlation coefficient (CC) by 6% relative to GLM for climatology mean precipitation. Especially, RRDBNet has substantial improvements in extreme precipitation compared with other models. It reduces RMSE by 58% (79%) and improves CC by 38% (145%) relative to GLM for R95P (R99P), where R95P and R99P represent extreme precipitation and very extreme precipitation, respectively. For the probability density function of daily precipitation, it is further demonstrated that RRDBNet performs better as regards extreme precipitation frequency. Our results suggest that statistical downscaling based on RRDBNet may be an effective tool for historical and future climate simulations from global climate models.

A Novel Hybrid Method for Predicting Specific Crop Yield

Predicting crop yields is crucial for economic planning, food security, and agricultural management. Conventional approaches frequently need accurate results regarding the complex interaction between environmental conditions and crop-specific traits. By combining the best features of machine learning algorithms with expert-level domain knowledge, we provide a new hybrid approach to predicting crop yields that is both state- and crop-specific. Our approach integrates historical yield data, satellite imagery, meteorological data, soil properties, and crop-specific information to build robust predictive models. Specifically, we employ ensemble learning techniques, including random forest and gradient boosting, to capture complex nonlinear relationships and improve prediction accuracy. Additionally, we incorporate domain knowledge through feature engineering and selection to enhance model interpretability and generalization. We validate our method using real-world datasets from diverse geographical regions and crops, demonstrating its superior performance compared to traditional approaches. Our proposed hybrid method offers a promising solution for accurate and reliable crop yield prediction, facilitating informed agricultural decision-making.

Machine learning predictions on the compressive stress–strain response of lattice-based metamaterials

Predicting the stress–strain curve of lattice-based metamaterials is crucial for their design and application. However, the complex nonlinear relationship between the mesoscopic structure of lattice materials and their macroscopic mechanical behavior makes prediction challenging. In this study, beam element models of over 20,000 lattice structures were established using Python scripts, and calculations were performed by ABAQUS to obtain training and testing datasets. The spatial features of each lattice-based metamaterial were then encoded into a graph, a data structure recognizable by machine learning algorithm. Utilizing machine learning methods, a Structure to Sequence Neural Network was constructed and trained, achieving rapid prediction of the compressive stress–strain curves for lattice-based metamaterials. Afterwards, several lattice structures were randomly selected and 3D printed. The accuracy of the simulation results as well as machine learning predictions was validated through quasi-static compression experiments. It is revealed that the proposed Neural Network model outperforms the traditional Artificial Neural Networks as the errors are reduced while the Coefficient of Determination is higher. The results demonstrate the accurate fitting between the complex spatial features of the lattice-based metamaterials and their stress–strain curves, which provides a potential methodology for inverse optimization of the lattice-based metamaterials in the future.

An ensemble spatial prediction method considering geospatial heterogeneity

Ensemble learning synthesizes the advantages of different models and has been widely applied in the field of spatial prediction. However, the nonlinear constraints of spatial heterogeneity on the model ensemble process make it difficult to adaptively determine the ensemble weights, greatly limiting the predictive ability of the ensemble learning model. This paper therefore proposes a novel geographical spatial heterogeneous ensemble learning method (GSH-EL). Firstly, the geographically weighted regression model, geographically optimal similarity model, and random forest model are used as three base learners to express local spatial heterogeneity, global feature correlation, and nonlinear relationship of geographic elements, respectively. Then, a spatially weighted ensemble neural network module (SWENN) of GSH-EL is proposed to express spatial heterogeneity by exploring the complex nonlinear relationship between the spatial proximity and ensemble weights. Finally, the outputs of the three base learners are combined with the spatial heterogeneous ensemble weights from SWENN to obtain the spatial prediction results. The proposed method is validated on the PM2.5 air quality and landslide dataset in China, both of which obtain more accurate prediction results than the existing ensemble learning strategies. The results confirm the need to accurately express spatial heterogeneity in the model ensemble process.

A Stock Index Futures Price Prediction Approach Based on the MULTI-GARCH-LSTM Mixed Model

As a type of financial derivative, the price fluctuation of futures is influenced by a multitude of factors, including macroeconomic conditions, policy changes, and market sentiment. The interaction of these factors makes the future trend become complex and difficult to predict. However, for investors, the ability to accurately predict the future trend of stock index futures price is directly related to the correctness of investment decisions and investment returns. Therefore, predicting the stock index futures market remains a leading and critical issue in the field of finance. To improve the accuracy of predicting stock index futures price, this paper introduces an innovative forecasting method by combining the strengths of Long Short-Term Memory (LSTM) networks and various Generalized Autoregressive Conditional Heteroskedasticity (GARCH)-family models namely, MULTI-GARCH-LSTM. This integrated approach is specifically designed to tackle the challenges posed by the nonstationary and nonlinear characteristics of stock index futures price series. This synergy not only enhances the model’s ability to capture a wide range of market behaviors but also significantly improves the precision of future price predictions, catering to the intricate nature of financial time series data. Initially, we extract insights into the volatility characteristics, such as the aggregation of volatility in futures closing prices, by formulating a model from the GARCH family. Subsequently, the LSTM model decodes the complex nonlinear relationships inherent in the futures price series and incorporates assimilated volatility characteristics to predict future prices. The efficacy of this model is validated by applying it to an authentic dataset of gold futures. The empirical findings demonstrate that the performance of our proposed MULTI-GARCH-LSTM hybrid model consistently surpasses that of the individual models, thereby confirming the model’s effectiveness and superior predictive capability.

Markov-Embedded Affinity Learning with Connectivity Constraints for Subspace Clustering

Subspace clustering algorithms have demonstrated remarkable success across diverse fields, including object segmentation, gene clustering, and recommendation systems. However, they often face challenges, such as omitting cluster information and the neglect of higher-order neighbor relationships within the data. To address these issues, a novel subspace clustering method named Markov-Embedded Affinity Learning with Connectivity Constraints for Subspace Clustering is proposed. This method seamlessly embeds Markov transition probability information into the self-expression, leveraging a fine-grained neighbor matrix to uncover latent data structures. This matrix preserves crucial high-order local information and complementary details, ensuring a comprehensive understanding of the data. To effectively handle complex nonlinear relationships, the method learns the underlying manifold structure from a cross-order local neighbor graph. Additionally, connectivity constraints are applied to the affinity matrix, enhancing the group structure and further improving the clustering performance. Extensive experiments demonstrate the superiority of this novel method over baseline approaches, validating its effectiveness and practical utility.

Design and implementation of student management system of integrated programmable device programming system

With the continuous growth and development of society, the reform of higher education is gradually taking shape, and colleges and universities are taking more and more responsibilities in promoting education. The main element of university management is the student management system, which is very important to the development of the university. Under the objective environment of colleges and universities seeking to expand the scale of running schools and rapid economic and social development, the current student management system has been unable to meet the various needs of contemporary students. The integration of programmable device programming systems offers a student management system distinct advantages in terms of reliability, flexibility, and user-friendly operation. This study focuses on developing an effective and affordable student management system by incorporating a programmable device programming system. To evaluate the system's performance, this paper suggests the utilization of a BP neural network, renowned for its high nonlinear approximation capabilities and effectiveness in handling complex nonlinear functional relationships. The experimental findings highlight a significant contribution, demonstrating that the system achieved a throughput of 180 times per second, with a maximum CPU utilization of 99%. Notably, the system's stability exceeded 95%, contrasting sharply with the traditional student management system's stability at a mere 65%. These results underscore the substantial contribution of the proposed system, showcasing its enhanced stability compared to conventional student management systems.

Residual Spatiotemporal Convolutional Neural Network Based on Multisource Fusion Data for Approaching Precipitation Forecasting

Approaching precipitation forecast refers to the prediction of precipitation within a short time scale, which is usually regarded as a spatiotemporal sequence prediction problem based on radar echo maps. However, due to its reliance on single-image prediction, it lacks good capture of sudden severe convective events and physical constraints, which may lead to prediction ambiguities and issues such as false alarms and missed alarms. Therefore, this study dynamically combines meteorological elements from surface observations with upper-air reanalysis data to establish complex nonlinear relationships among meteorological variables based on multisource data. We design a Residual Spatiotemporal Convolutional Network (ResSTConvNet) specifically for this purpose. In this model, data fusion is achieved through the channel attention mechanism, which assigns weights to different channels. Feature extraction is conducted through simultaneous three-dimensional and two-dimensional convolution operations using a pure convolutional structure, allowing the learning of spatiotemporal feature information. Finally, feature fitting is accomplished through residual connections, enhancing the model’s predictive capability. Furthermore, we evaluate the performance of our model in 0–3 h forecasting. The results show that compared with baseline methods, this network exhibits significantly better performance in predicting heavy rainfall. Moreover, as the forecast lead time increases, the spatial features of the forecast results from our network are richer than those of other baseline models, leading to more accurate predictions of precipitation intensity and coverage area.

#960 Trends in eGFR and ultra-processed food intake in relationship to income among United States adults

Abstract Background and Aims Ultra-processed foods (UPF) are formulations of ingredients that are mostly of exclusive industrial use and may contain additives like artificial colors, flavors, or stabilizers. The sale and consumption of these foods has been increasing despite their associations with increased risk for several non-communicable diseases including chronic kidney disease. Ultra-processed foods are widely accessible, more shelf-stable and convenient while tending to be more affordable and lower in nutritional quality. The aim of this study was to evaluate trends in kidney function and UPF intake and temporal effect of income among United States (US) adults. Method We used dietary intake data collected with 24-hour dietary recalls from 35160 US adults across 8 cycles of the National Health and Nutrition Examination Survey (NHANES) (from 2003-2004 to 2017-2018 cycles) to categorize intake using the Nova processed food classification system. Poisson regression models were used to assess the 15-year trends in populations at risk for kidney function decline (measured by estimated glomerular filtration rate [eGFR]) and higher percentage of calories from UPF (PC-UPF) consumed along with the temporal effect of income in select subpopulations and adjusted for age and gender and non-linear effects of income and age). Results Over the past 15 years eGFR, PC-UPF, and self-reported income continue to increase among US adults. However, the associations between eGFR, PC-UPF and income are quite complex, vary by subpopulation and non-linearly depend on age. Specifically, among Black adults adjusted for age (age2) and gender, eGFR was the lowest (85 mL/min/1.73 m2, 95% CI: 84.9, 85.0 mL/min/1.73 m2) among those with the lowest ($2,500) and highest ($125,000) income (p = 0.0000000579 for a quadratic term of income), respectively. Hispanic and Mexican Americans also have similar pattern eGFR with the lowest and highest income though the eGFR is consistently higher than White and Black Americans. Across cycles, White Americans with the lowest income ($2,500) have the highest PC-UPF (range: 57-62%, p &amp;lt;2e-16; mean: 59.4%, 95% CI: 59.3, 59.6%) while Black Americans with the lowest ($2,500) and highest income ($125,000) have the highest PC-UPF (range: 53-58%, p &amp;lt;2e-16; mean: 54.8%, 95% CI: 54.3, 55.4%). Income levels does not predict PC-UPF among Mexican Americans. Overall, the PC-UPF is steadily increasing in all population subgroups. Conclusion Among Non-White US adults, higher eGFR was associated with higher income until reaching a group-specific threshold when the relationship reversed. We observed a similar threshold pattern in the association between PC-UPF and income among Black and Hispanic Americans. These complex non-linear relationships between eGFR, PC-UPF and income are driven in part by age-related association and warrant further exploration.

#2784 Deep learning algorithm predicts individual ultrafiltration trajectories for patients on automated peritoneal dialysis 12 weeks ahead

Abstract Background and Aims Adequate small solute clearance and ultrafiltration (UF) are the main requirements for fluid and salt homeostasis on peritoneal dialysis (PD). New automated peritoneal dialysis (APD) systems that allow remote patient monitoring in terms of PD treatment analysis have the potential to improve clinical outcomes. UF on APD although dependent on PD prescription, i.e. fill volumes and glucose concentration, exhibits substantial intra- and interindividual variation, even day-by-day. As conventional statistical methods are not able to assess these complex non-linear relationships adequately, we aimed to build a deep learning algorithm to predict ultrafiltration trajectories of individual patients 12 weeks ahead. Method Data on daily UF and PD prescriptions (i.e., glucose concentration, fill volumes) was extracted from the APD cycler management software (PD Link, Baxter, IL, USA) and patient charts at the Medical University of Vienna, for a secondary analysis of this cohort. The Keras high-level API for TensorFlow in R software (R Core Team 2023) was used to create a Long Short-Term Memory (LSTM) recurrent neural network based on historic prescription and gcUF data of 26 weeks (after the first 90 days of APD) to predict individual UF trajectories for the following 12 weeks. Results APD machine readouts from a total of 123 patients, with a mean time on PD of 117 (±73) days, were retrieved. The final recurrent neural network consisted of four LSTM layers with 64 cells each, sigmoid activation. The final model was fit using 30 epochs, 49 steps per epoch, and recurrent dropout to avoid overfitting. A 40-40-20% data split was used to train, validate, and test the model adequately. This deep learning algorithm was able to predict individual UF values, for each day of the upcoming 12 weeks, with an accuracy of at mean 201ml. Upon visual inspection, predicted individual UF trajectories smoothly resemble the overall day-by-day trajectory of the patient but fail to account for extreme spikes in either direction. Conclusion In conclusion, historic daily APD cycler and prescription data in coalescence with respective daily measurements of membrane function (i.e., gcUF) can be used to predict individual daily ultrafiltration values and trajectories of APD patients 12 weeks ahead using a deep learning algorithm. Therefore, continuous collection of PD prescription data and UF measurements may be used to monitor peritoneal membrane function of patients on APD and alert clinical teams upon detection of relevant changes of UF trajectories already 12 weeks ahead to facilitate early intervention. In contrast to current “screening” methods on PD, i.e., Peritoneal Equilibrium Test, these methods could reduce treatment burden.

MvMRL: a multi-view molecular representation learning method for molecular property prediction.

Effective molecular representation learning is very important for Artificial Intelligence-driven Drug Design because it affects the accuracy and efficiency of molecular property prediction and other molecular modeling relevant tasks. However, previous molecular representation learning studies often suffer from limitations, such as over-reliance on a single molecular representation, failure to fully capture both local and global information in molecular structure, and ineffective integration of multiscale features from different molecular representations. These limitations restrict the complete and accurate representation of molecular structure and properties, ultimately impacting the accuracy of predicting molecular properties. To this end, we propose a novel multi-view molecular representation learning method called MvMRL, which can incorporate feature information from multiple molecular representations and capture both local and global information from different views well, thus improving molecular property prediction. Specifically, MvMRL consists of four parts: a multiscale CNN-SE Simplified Molecular Input Line Entry System (SMILES) learning component and a multiscale Graph Neural Network encoder to extract local feature information and global feature information from the SMILES view and the molecular graph view, respectively; a Multi-Layer Perceptron network to capture complex non-linear relationship features from the molecular fingerprint view; and a dual cross-attention component to fuse feature information on the multi-views deeply for predicting molecular properties. We evaluate the performance of MvMRL on 11 benchmark datasets, and experimental results show that MvMRL outperforms state-of-the-art methods, indicating its rationality and effectiveness in molecular property prediction. The source code of MvMRL was released in https://github.com/jedison-github/MvMRL.

Conditional generative adversarial networks for individualized causal mediation analysis

Abstract Most classical methods popularly used in causal mediation analysis can only estimate the average causal effects and are difficult to apply to precision medicine. Although identifying heterogeneous causal effects has received some attention, the causal effects are explored using the assumptive parametric models with limited model flexibility and analytic power. Recently, machine learning is becoming a major tool for accurately estimating individualized causal effects, thanks to its flexibility in model forms and efficiency in capturing complex nonlinear relationships. In this article, we propose a novel method, conditional generative adversarial network (CGAN) for individualized causal mediation analysis (CGAN-ICMA), to infer individualized causal effects based on the CGAN framework. Simulation studies show that CGAN-ICMA outperforms five other state-of-the-art methods, including linear regression, k-nearest neighbor, support vector machine regression, decision tree, and random forest regression. The proposed model is then applied to a study on the Alzheimer’s disease neuroimaging initiative dataset. The application further demonstrates the utility of the proposed method in estimating the individualized causal effects of the apolipoprotein E-ε4 allele on cognitive impairment directly or through mediators.