Data-driven Prediction Research Articles

Prediction intervals with controlled length in the heteroscedastic Gaussian regression

We tackle the problem of building a prediction interval in heteroscedastic Gaussian regression. We focus on prediction intervals with constrained expected length in order to guarantee interpretability of the output. In this framework, we derive a closed-form expression of the optimal prediction interval that allows for the development of a data-driven prediction interval based on plug-in. The construction of the proposed algorithm is based on two samples, one labelled and another unlabelled. Under mild conditions, we show that our procedure is asymptotically as good as the optimal prediction interval both in terms of expected length and error rate. In particular, the control of the expected length is distribution-free. We also derive rates of convergence under smoothness and the Tsybakov noise conditions. We conduct a numerical analysis that exhibits the good performance of our method. It also indicates that even with a few amount of unlabelled data, our method is very effective in enforcing the length constraint.

Data-Driven Prediction of COF in Wet Friction Components: A Model Development and Interpretability Analysis

Abstract Predicting the coefficient of friction (COF) is essential for enhancing the efficiency and reliability of mechanical systems. Nevertheless, traditional mechanistic models relying on fixed values or fitted curves fail to accurately capture this complexity. To address this issue, this paper proposes a model for predicting the COF of wet friction components using an extreme gradient boosting (XGBoost) algorithm optimized by the sparrow search algorithm (SSA). This model effectively captures the nonlinear relationships among relative speed, pressure, temperature, and COF. As a result, the proposed SSA-XGBoost model exhibits excellent predictive performance with an RMSE of only 0.063, and 88.3% of the COF predictions have a relative error of less than 1%, significantly outperforming other deep learning algorithms. Additionally, to enhance the understanding of the COF prediction results for wet friction components, SHAP interpretability analysis is used to explore the influence of relative speed, pressure, and temperature on the predicted COF values.

A Multi-Objective Energy Optimization Model for Renewable Energy Systems with Hybrid Energy Storage

Building a Renewable Energy System (RES) is a viable way to address resource depletion and accomplish decarbonization. This study presents a hybrid power system that combines wind, solar, battery, and thermal energy storage. It also examines the co-optimization of scheduling and operation for multiple objectives. The hybrid system combines the affordability of thermal energy storage (TES) with the adaptability of batteries to effectively address the issue of intermittent RES. A new approach for integrated operation is offered, which relies on the electrical block's operation limit. The planning-operation co-optimization system considers minimizing the Net Present Cost (NPC) and reducing power supply likelihood todetermine the best operation limit and sizing decision factors. The co-optimization issue is addressed using novel multi-objective decision-making (MO-DM). This approach incorporates the Decision-Maker (DM) preferences data to direct the evolutionary process toward the desired area. In addition, a data-driven prediction model captures the risks and losses associated with wind generation. The case study's findings indicate the following: (1) The data-driven system has a higher level of precision in wind power projection when compared with commonly used physical simulations. (2) The suggested MO-DM exhibits superior integration, variety, and robustness achievement in the DMchosen area compared to others. (3) The combined battery-thermal energy storing structure achieves improved economy and dependability through the optimum organized operation approach, in contrast to using a single energy storage system under various testing constraints.

A generalized machine learning framework for data-driven prediction of relative density in laser powder bed fusion parts

A generalized machine learning framework for data-driven prediction of relative density in laser powder bed fusion parts

Prediction of shield tunneling attitudes: A muti-dimensional feature synthesizing and screening method

Shield attitudes, essentially governed by intricate mechanisms, impact the segment assembly quality and tunnel axis deviation. In data-driven prediction, however, existing methods using the original driving parameters fail to present convincing performance due to insufficient consideration of complicated interactions among the parameters. Therefore, a multi-dimensional feature synthesizing and screening method is proposed to explore the optimal features that can better reflect the physical mechanism in predicting shield tunneling attitudes. Features embedded with physical knowledge were synthesized from seven dimensions, which were validated by the clustering quality of Shapley Additive Explanations (SHAP) values. Subsequently, a novel index, Expected Impact Index (EII), has been proposed for screening the optimal features reliably. Finally, a Bayesian-optimized deep learning model was established to validate the proposed method in a case study. Results show that the proposed method effectively identifies the optimal parameters for shield attitude prediction, with an average Mean Squared Error (MSE) deduction of 27.3%. The proposed method realized effective assimilation of shield driving data with physical mechanism, providing a valuable reference for shield deviation control.

Data-driven prediction of flow fields in a needle-ring-net electrohydrodynamic pump system

In this paper, a data-mechanism hybrid modeling method for efficiently obtaining an electrohydrodynamic flow field is proposed. First, a backpropagation (BP) model with high accuracy is trained to get the value of essential parameter q0 for the mechanism simulation of flow fields. Subsequently, the mechanism model is used to generate a database for flow field reconstruction. Three machine learning algorithms, namely, BP neural network, random forest regression (RFR), and convolutional neural network (CNN), are employed to predict and reconstruct the flow behaviors of a needle-ring-net electrohydrodynamic pump. The RFR model demonstrates higher accuracy and precision in predicting velocity and pressure in the flow field compared to the BP and CNN models. The use of machine learning models for flow field prediction can significantly reduce the computational time while maintaining the computational accuracy. Additionally, an analysis assessing the impact of varying dataset sizes on the prediction accuracy of the model is conducted. The results indicate that the size of the dataset significantly influences the model predictive performance. Specifically, larger datasets are suggested to enhance both the accuracy and the generalization capabilities of the model. This observation highlights the critical role of dataset size in optimizing the performance of machine learning models for predictive tasks in engineering applications. These results offer important references for improving the design and optimization of electrohydrodynamic pumps.

Data-driven prediction of higher-order structure in carbon nanotube films for high thermoelectric power factor

We optimized the higher-order structures and semiconducting purity of single-walled carbon nanotubes (SWCNTs) to enhance the thermoelectric power factor PF by combining the thermoelectric random stick network (TE-RSN) method and a genetic algorithm. The PF of the optimized films was increased approximately fivefold for initial random structures. In addition, while the random structures showed the maximum PF when the ratio of semiconducting to metallic SWCNTs R s exceeded 0.98, the optimized structures converged to an R s of approximately 0.9. The optimized structures exhibited an increased local density and the peak of alignment angle distribution, leading to an increase in both the electrical conductivity and the Seebeck coefficient. We interpreted the increase in the Seebeck coefficient using a serial model. The results indicated that the reduction in the number of contacts within the paths and the subsequent increase in temperature difference on semiconducting SWCNTs led to the increase in the Seebeck coefficient.

Data-driven prediction of early language delay at Youth Health Care Services in the Netherlands

Abstract Introduction Language development in early childhood significantly impacts future academic success, social interaction, and emotional well-being. This study aimed to develop and evaluate a prediction model using data on developmental milestones and parents’ socio-economic background known at age 2, and translate the results into an interactive digital dashboard and pilot it with Youth Health Care (YHC) professionals and parents. Methods A mixed-method study on retrospective cohort of children attending YHC at age 2, followed up until age 4. Data on language development, gender, parental education, and home language environment were obtained from YHC digital dossiers and linked to parental socio-economic status characteristics, and perinatal outcomes and healthcare costs using a secure remote access environment of Statistics Netherlands. Random forest and logistic regression models were computed with language development at age 4 as an outcome, initially using all available data at age 2 as predictors and then with a restricted model using variables readily available in YHC settings. Models performance was assessed using sensitivity, specificity, and AUC value. Results were visualized through an interactive dashboard and pilot-tested in simulated consultations. Results Among 9,148 children, 13.2% had language development delay at age 4. The full model had an AUC of 0.78, while the restricted model (gender, parental education, language environment, and language development), achieved an AUC of 0.77. YHC professionals and parents recognized the value of individualized, data-driven risk assessment for potential language delay to stimulate the discussion of preventive interventions. Conclusions A good quality prediction model for language development at age 4 can be derived from just a few background characteristics of the child and parents at age 2. The digital dashboard presents a practical approach of integrating prediction model results into daily YHC practice. Key messages • Language development at age of 4 can be predicted as early as age of 2 using just a few background characteristics of child and parents, which presents a window of opportunity for prevention. • The digital dashboard presents a practical approach of integrating prediction model results into preventive care practice.

Data-driven Prediction of Underwater Navigation Fitness Zone Classification

This article presents a pioneering approach to enhancing underwater navigation accuracy and stealth through the utilization of gravity anomaly data. Developed by researchers from leading universities in China, the study aims to revolutionize autonomous marine navigation systems in unknown sea areas. By integrating advanced data processing, feature extraction, model construction, and optimization techniques, a cutting-edge navigation system has been formulated, showcasing remarkable improvements in navigation precision.The core innovation lies in leveraging bicubic interpolation to refine data resolution, enabling a meticulous evaluation of regional suitability and segmentation of sea zones. Subsequently, a gradient-based suitability prediction model is constructed, which is validated through migration prediction, underscoring its effectiveness and reliability. This research not only addresses the limitations of traditional terrain matching navigation methods, plagued by data accuracy and resolution constraints, but also paves the way for advancements in oceanic exploration and technological competitiveness.The introduction of gravity anomaly data, coupled with machine learning and visualization tools, marks a significant step forward in the development of adaptive navigation zones. This innovative system holds immense potential to enhance the autonomy, accuracy, and concealment of underwater vehicles, thereby facilitating safer and more efficient oceanic missions in the future.

Data-driven prediction of the mechanical behavior of nanocrystalline graphene using a deep convolutional neural network with PCA

Data-driven prediction of the mechanical behavior of nanocrystalline graphene using a deep convolutional neural network with PCA

Vortex-Induced Vibration Performance Analysis of Long-Span Sea-Crossing Bridges Using Unsupervised Clustering

Vortex-induced vibration is a type of wind-induced vibration occurring frequently in large-span sea-crossing bridges under relatively low wind speeds, posing a threat to the structural fatigue performance and driving comfort. Identifying the instantaneous occurrence moments of vortex-induced vibration is a prerequisite for establishing a data-driven prediction model for vortex-induced vibration, and it is of great significance for the monitoring and early warning of vortex-induced vibration performance in bridges. To automatically detect the occurrence moments of vortex-induced vibration and establish a correlation model between vortex-induced vibration amplitude and environmental factors, this study proposes a fuzzy C-means clustering-based classification method. In order to detect the occurrence moments of vortex-induced vibration more finely, only short-term or even instantaneous structural vibration indicators were selected and transformed for distribution as clustering features. The entire detection process could be carried out unsupervised, reducing the manual cost of obtaining vortex-induced vibration information from massive monitoring data. Finally, actual vortex-induced vibration test data from a certain overseas bridge was utilized to verify the feasibility of this method. Based on the classification results, the correlation between vortex-induced vibration amplitude and environmental variables was determined, providing valuable guidance for predicting vortex-induced vibration amplitudes.

Data-Driven Prediction of Nitrogen Stress and Crop Yield in a Maize -Wheat Cropping System

The study has attempted to determine leaf-N content in a cropping system consisting of Wheat and Maize using plant sensors like GreenSeeker and SPAD meter (chlorophyll meter) and use machine learning models so that crop yields based on inputs can be predicted of data on nitrogen doses, NDVI values, SPAD values, and leaf N content. The data recorded from field experiments on wheat and maize crop variable fertilized by N (wheat: 0, 30, 60, 90, 120, 150, 180 and 240 kg N/ha; maize: 40, 80, 120, 160, 200, 240 and 300 kg N/ha) were used. A pair plot analysis showed a positive association between leaf N and NDVI, between leaf N and SPAD values, and between NDVI and SPAD values; all these parameters had a positive correlation with N- application rates. For yield prediction, the model calibrated for wheat could validate yield prediction for maize as well. Random Forest and Support Vector Machines showed reproducible performance (consistent) on both datasets (maize and wheat), with accuracies of 85.71% and 100%, respectively, on the original dataset.

A data-driven prediction for concrete crack propagation path based on deep learning method

Accurate and efficient prediction of concrete crack propagation path facilitates timely maintenance. To avoid errors arising from assumptions in physical models and parameter measurements, a data-driven prediction method for concrete crack propagation path is proposed. The spatio-temporal prediction approach is presented by unifying the conditional Generative Adversarial Network and Long Short-Term Memory Network (UcGLS-Net). The UcGLS-Net has the advantage of enabling temporal (by Long Short-Term Memory Network, LSTM) and spatial information (by the conditional Generative Adversarial Network, cGAN) in the concrete crack propagation process to be treated simultaneously and distinctly thus improving prediction accuracy. Notably, results demonstrate that UcGLS-Net still works in crack propagation path prediction even with unknown aggregate distribution. Moreover, the effects of historical input quantity and prediction intervals on prediction accuracy of the UcGLS-Net are studied and the network framework is optimized. Due to high efficiency, this study provides a new adaptable and real-time prediction method for concrete crack propagation path, offering support for the design, damage prediction, assessment, and maintenance of concrete structures.

Multimodal data-driven, vertical visualization prediction model for early prediction of atherosclerotic cardiovascular disease in patients with new-onset hypertension.

: Hypertension is an important contributing factor to atherosclerotic cardiovascular disease (ASCVD), and multiple risk factors, many of which are implicated in metabolic disorders, contribute to the cause of hypertension. Despite the promise of multimodal data-driven prediction model, no such prediction model was available to predict the risk of ASCVD in Chinese individuals with new-onset hypertension and no history of ASCVD. : A total of 514 patients were randomly allocated to training and verification cohorts (ratio, 7 : 3). We employed Boruta feature selection and conducted multivariate Cox regression analyses to identify variables associated with ASCVD in these patients, which were subsequently utilized for constructing the predictive model. The performance of prediction model was assessed in terms of discriminatory power (C-index), calibration (calibration curves), and clinical utility [decision curve analysis (DCA)]. : This model was derived from four clinical variables: 24-h SBP coefficient of variation, 24-h DBP coefficient of variation, urea nitrogen and the triglyceride-glucose (TyG) index. Bootstrapping with 500 iterations was conducted to adjust the C-indexes were C-index = 0.731, 95% confidence interval (CI) 0.620-0.794 and C-index: 0.799, 95% CI 0.677-0.892 in the training and verification cohorts, respectively. Calibration plots with 500 bootstrapping iterations exhibited a strong correlation between the predicted and observed occurrences of ASCVD in both the training and verification cohorts. DCA analysis confirmed the clinical utility of this prediction model. The constructed nomogram demonstrated significant additional prognostic utility for ASCVD, as evidenced by improvements in the C-index, net reclassification improvement, integrated discrimination improvement, and DCA compared with the overall ASCVD risk assessment. The developed longitudinal prediction model based on multimodal data can effectively predict ASCVD risk in individuals with an initial diagnosis of hypertension. : The trial was registered in the Chinese Clinical Trial Registry (ChiCTR2300074392).

Maximizing the spread of information through content optimization

As data-driven prediction models advance, an increasing number of people are enjoying news personalized to their interests. The primary problem such recommendation models solve is to precisely match information with users and, in so doing, ensure that news spreads with greater efficiency. However, these techniques only help the media platform; they do not help those who produce the news. Hence, we devised a propagation framework based on a human-in-the-loop simulation that helps content authors maximize the spread of their messages through social networks. The framework works by acting on feedback provided by the simulation model. Additionally, the spread of information is formulated as a multi-objective optimization problem in which propagation is data-driven and simulated with machine learning techniques that leverage data on the historical behaviors of users. We additionally describe an implementation for this framework as an example of how the framework might be used in real life. On the practical side, the implementation uses text data from a blog to simulate the message's propagation, while, from a technical point of view, the multi-objective optimization problem is divided into an information retrieval problem and an integer programming problem, the results of which are fed back into the content editor as content operation strategies. A case study with the Sina Weibo microblog site not only validates the framework but also provides practitioners with insights into how to maximize the spread of information through social networking platforms. The results show that the proposed propagation framework is capable of increasing retweets by 7.9575 %. As an interesting aside, our experiments also show that the Weibo retweet lottery is both popular and a highly effective mechanism for increasing reposts.

Analysis of Weighted Factors Influencing Submarine Cable Laying Depth Using Random Forest Method

This study addresses the limitations of traditional methods used to analyze factors influencing submarine cable burial depth and emphasizes the underutilization of cable construction data. To overcome these limitations, a machine learning-based model is proposed. The model utilizes cable construction data from the East China Sea to predict the weight of factors influencing cable burial depth. Pearson correlation analysis and principal component analysis are initially employed to eliminate feature correlations. The random forest method is then used to determine the weights of factors, followed by the construction of an optimized backpropagation (BP) neural network using the ISOA-BP hybrid optimization algorithm. The model’s performance is compared with other machine learning algorithms, including support vector regression, decision tree, gradient decision tree, and the BP network before optimization. The results show that the random forest method effectively quantifies the impact of each factor, with water depth, cable length, deviation, geographic coordinates, and cable laying tension as the significant factors. The constructed ISOA-BP model achieves higher prediction accuracy than traditional algorithms, demonstrating its potential for quality control in cable laying construction and data-driven prediction of cable burial depth. This research provides valuable theoretical and practical implications in the field.

Data-driven prediction of drilling strength ahead of the bit

This paper compares the performance of two data-driven methods, Signal-Matching Predictor (SMP) and Long Short-Term Memory (LSTM), for predicting drilling strength (Es) ahead of the bit based on drilling data from nearby offset wells. The comparison is based on the accuracy, applicability, complexity, and computational cost of the methods with the objective of suggesting the most appropriate tool for look-ahead drilling strength prediction. The methods were tested using data from offshore wells in Newfoundland. The SMP used a fixed-size sliding-window of real-time Es data from the target well to find a match in the offset well within similar geological formations and chose the scaled value from the offset well as the prediction. In the second approach, twelve LSTM models were trained using the drilling data of twelve offset wells, and the drilling data of the thirteenth well was used for blind testing. Results showed that the SMP achieved a coefficient of determination (R2) of 0.92, 0.92, and 0.79 for predicting 1.5, 3, and 5 feet ahead of the bit, respectively, while the LSTM reached an R2 of 0.95, 0.92, and 0.80 for the respective prediction intervals. The R2 of the LSTM models was further increased to 0.96, 0.94, and 0.83 after retraining it with weighted samples in formation transition zones. Also, a post-processing technique was proposed that further enhanced the R2 of the LSTM-based approach to 0.98, 0.97, and 0.93, respectively. The strength of the LSTM-based approach was to use measurable drilling parameters as the only inputs and not the Es itself. According to the results, the LSTM-based method can be reliably used to predict the Es ahead of the bit allowing drillers to identify upcoming drilling dysfunctions.

The static voltage stability analysis of photovoltaic energy storage systems based on NPU algorithm

Although the data-driven static voltage stability problems have been widely studied, most of the classical algorithms focus more on improving the accuracy of the system prediction, ignoring the error classification errors generated during the prediction process. Furthermore, current research ignores the utilization of data-driven voltage stability assessment of energy storage systems. Therefore, this paper proposes a static voltage stability assessment method for photovoltaic energy storage systems based on considering the error classification constraint algorithm using Neyman-Pearson umbrella algorithms. Firstly, the Spearman Correlation Coefficient is employed in the feature selection phase. Secondly, an updated voltage stability assessment (VSA) model is proposed. Compared with the existing data-driven prediction of system static voltage stability in the literature, it can realize voltage stability assessment more quickly. Furthermore, on the basis of rapid voltage stability assessment, the umbrella NP classifier can also effectively limit the first-class error and attenuate the effect of error classification by mirroring the control of the number of cycle splits and the type I classification error threshold. Finally, the simulation and experimental results show that the effectiveness and robustness of the scheme proposed in this paper in grid-connected photovoltaic energy farms.

Data-driven energy utilization for plug-in hybrid electric bus with driving patten application and battery health considerations

The efficient utilization of energy is a crucial consideration for plug-in hybrid electric buses (PHEB). However, achieving the optimal energy management strategy (EMS) for PHEB necessitates the harmonious optimization of both driving conditions and battery status. Consequently, an innovative EMS that integrates the data-driven prediction of the driving cycles and battery health status. Explicitly, the bi-directional long short-term memory algorithm is employed to reconstruct and train the health status data of power batteries, enabling the accurate estimation of their State of Health (SOH) values. Moreover, a database of historical driving behavior is meticulously collected through a real-world PHEB platform, and the driving cycles for a certain while are predicted with the clustering algorithm. Then, this invaluable information on the battery status and driving cycles is seamlessly introduced into an adaptive-network-based fuzzy inference system (ANFIS) to regulate of the energy conversion factor in the following equivalent consumption minimization strategy (ECMS). Besides, the offline global optimization of equivalent fuel consumption with the Pontryagin's minimum principle (PMP) is preliminarily conducted to generate the pre-established energy conversion factors, facilitating convenient access to the engine and motor's reasonable power distribution range. Finally, simulations and experiments have validated the effectiveness of the proposed EMS, demonstrating a 23.57 % reduction in battery SOC depletion and a 10.53 % decrease in fuel consumption compared to traditional ECMS methods. More importantly, the proposed method has been validated for its executability in embedded devices through real-world bench experiments. This research holds significant reference value for energy utilization and practical implementation in the PHEB, thereby contributing to the advancement of knowledge in this field.

Data-driven prediction of scaling and ignition of inertial confinement fusion experiments

Recent advances in inertial confinement fusion (ICF) at the National Ignition Facility (NIF), including ignition and energy gain, are enabled by a close coupling between experiments and high-fidelity simulations. Neither simulations nor experiments can fully constrain the behavior of ICF implosions on their own, meaning pre- and postshot simulation studies must incorporate experimental data to be reliable. Linking past data with simulations to make predictions for upcoming designs and quantifying the uncertainty in those predictions has been an ongoing challenge in ICF research. We have developed a data-driven approach to prediction and uncertainty quantification that combines large ensembles of simulations with Bayesian inference and deep learning. The approach builds a predictive model for the statistical distribution of key performance parameters, which is jointly informed by past experiments and physics simulations. The prediction distribution captures the impact of experimental uncertainty, expert priors, design changes, and shot-to-shot variations. We have used this new capability to predict a 10× increase in ignition probability between Hybrid-E shots driven with 2.05 MJ compared to 1.9 MJ, and validated our predictions against subsequent experiments. We describe our new Bayesian postshot and prediction capabilities, discuss their application to NIF ignition and validate the results, and finally investigate the impact of data sparsity on our prediction results.