Search

A robust short-term significant wave height (Hs) modelling framework based on an ensemble local mean decomposition method integrated with random forest (i.e. , En-RLMD-RF) is developed. The robust local mean decomposition (RLMD) decomposed the Hs data series into three subseries; amplitude modulation, frequency modulation and the low-frequency product function (PFs). The partial autocorrelation function was employed to determine the correlation-based significant predictor signals between the PFs at t 0 and t 1. Then the statistically significant PFs were incorporated into the random forest (RF) to construct the RLMD-RF model. The RLMD-RF based forecasted PFs were used again in the RF model as input predictors resulting in an ensemble-based RLMD-RF (i.e. , En-RLMD-RF) model for forecasting short-term Hs. The En-RLMD-RF model is validated and compared with RF, extreme learning machine (ELM) and multiple linear regression (MLR) models and their hybrids RLMD-RF, RLMD-ELM, RLMD-MLR, En-RLMD-ELM and En-RLMD-MLR counterparts using a set of performance metrics. The results demonstrated that the En-RLMD-RF model generates better forecasting accuracy against the benchmarking models. This study is beneficial for the application and optimization of more clean energy resources worldwide for sustained energy generation. [ABSTRACT FROM AUTHOR]

• A spatiotemporal risk prediction of wildlife-vehicle collisions (WVCs) is presented. • A Bavarian dataset of around 731,000 accidents over a period of 10 years (2010–2019) was used. • Heterogenous datasets of impact factors are pre-processed using GIS. • Different machine learning (ML) approaches are applied - random forest, feedforward neural networks, and support vector machine classifier to predict the risk of WVCs. • ML algorithms provide high risk prediction quality of over 86% (double-weighted sensitivity and single-weighted specificity rate). • Spatial transferability and temporal forecast of WVC prediction was verified. Introduction: The technology in the automotive industry is becoming increasingly safer in the age of automated driving, but the number of accidents is still high, especially in wildlife-vehicle collisions (WVCs). To better avoid these accidents, patterns involved in these accidents must be detected. Method : This paper presents a spatiotemporal risk prediction of WVCs, including various road and environmental characteristics. A process of data preparation using GIS automated by Python scripts was developed to enable a spatiotemporal link of diverse features for the subsequent predictive data analysis. Different machine learning (ML) approaches were applied- random forest (RF), feedforward neural networks (FNN), and support vector machine classifier (SVM) - including automated ML to predict the risk of WVCs. Therefore, a dataset of approximately 731,000 accidents reported to the police in Bavaria over a period of 10 years (2010–2019) was used. In addition, non-accidents were randomly generated in Python over time and space for the supervised ML processes. As the actual risk probability for WVCs and non-WVCs is not entirely known, the impact of different training ratios between accidents and non-accidents was tested on the risk prediction quality (RPQ) (25%, 50%, 75%, 90% WVCs) of the double-weighted sensitivity and single-weighted specificity rate. Results: The test yielded high mean values of RPQ as an indicator for a suitable WVC prediction. Both RF (86.6%) and FNN (86.7%) were identified as suitable choices for WVC risk prediction in terms of RPQ. The SVM yielded a lower prediction quality, even though acceptable results could be achieved within a shorter runtime. Practical Applications: Spatial transferability was verified since the algorithm was trained on the dataset of Bavaria (excluding Upper Bavaria) and successfully tested in Upper Bavaria. WVC forecasts were also proven through training with datasets from 2010-2017 and in prediction for 2018 and 2019. [ABSTRACT FROM AUTHOR]

Accurate and reliable prediction of photovoltaic (PV) power output is critical to electricity grid stability and power dispatching capabilities. However, PV power generation is highly volatile and unstable due to different reasons. The wavelet transform (WT) has been utilised in time series applications, such as PV power prediction, to model the stochastic volatility and reduce prediction errors. Yet the existing WT approach has a limitation in terms of time complexity. It requires reconstructing the decomposed components and modelling them separately and thus needs more time for reconstruction, model configuration and training. The aim of this study is to improve the efficiency of applying WT by proposing a new method that uses a single simplified model. Given a time series and its WT coefficients, it trains one model with the coefficients as features and the original time series as labels. This eliminates the need for component reconstruction and training numerous models. This work contributes to the day-ahead aggregated solar PV power time series prediction problem by proposing and comprehensively evaluating a new approach of employing WT. The proposed approach is evaluated using 17 months of aggregated solar PV power data from two real-world datasets. The evaluation includes the use of a variety of prediction models, including Linear Regression, Random Forest, Support Vector Regression, and Convolutional Neural Networks. The results indicate that using a coefficients-based strategy can give predictions that are comparable to those obtained using the components-based approach while requiring fewer models and less computational time. [ABSTRACT FROM AUTHOR]

Since the frequency and intensity of heatwaves are expected to increase due to global warming, it is crucial to establish a simplified approach to preventing personal heat-related illnesses as occupational hazard under extremely hot environments, by taking into account individual differences in heat strain. In light of this, this study proposed a simplified forecast model for personal heat strain under extremely hot environments by utilizing feature importance in machine learning, focused on enhancing its field applicability. Using the 1417 records gathered from the experiments under extremely hot conditions, the simplified model was developed by utilizing four types of machine learning algorithms (i.e., random forest, extreme gradient boosting, support vector regression, and multi-layer perceptron). As a result, the proposed simplified models with four principal features (i.e., dry-bulb temperature, radiant temperature, relative humidity, and percentage body fat) were found to be the most reliable, resulting in a small difference of 0.047 °C compared to the reference model. Since the principal biometric features can be easily measured by a simple and low-cost body composition test (but not with the physical measurement of personal core body temperature) before they are put into the field, the proposed approach will aid in providing more systematic and efficient management system for proactively preventing personal heat-related illnesses. Furthermore, it is expected that the field applicability of the proposed approach can be continuously enhanced by utilizing the accumulated big data gathered from the field-based living-lab projects over a lengthy period. • Traditional approaches to heat stress primarily focused on environmental factors. • Recognizing individual differences in heat strain as an occupational hazard is imperative. • Principal features affecting personal heat strain were identified based on feature importance. • A simplified forecast model for personal heat strain was proposed using principal features. • The proposed model exhibited a small difference of 0.047 °C compared to the reference model. [ABSTRACT FROM AUTHOR]

The ability to predict global horizontal irradiance (GHI) in the wake of changing weather conditions is becoming increasingly important in driving economic growth in the renewable energy industry. The choice of relevant meteorological variables for predicting GHI is crucial because it influences the prediction accuracy due to various geographical conditions. To fulfill this aim, this proposed study identifies a subset of the relevant input variables for the prediction of GHI by applying two methods Feature Combination (FC) and Feature Selection (FS). The results reveal that the most significant input variables for predicting GHI are Solar Zenith Angle, Dew Point, Diffuse Horizontal irradiance, Direct Normal Irradiance, and Wind speed obtained by the FS method. The predictive performance of the selected features is evaluated by feeding them into six different types of Machine Learning (ML) regressor algorithms such as Multiple Linear regressor (MLR), Decision Tree (DT), Random Forest (RF), Gradient Boost (GB), Light Gradient Boost Machine (LGBM) and Extra Tree (ET). The performances of models are evaluated by using statistical measures such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). ET regressor gives the best prediction performance among all the six models. It is observed that the percentage reduction in error for the FS method as compared to the FC method in terms of the statistical indicators MAE, RMSE, and MAPE are 8.5%, 7.72%, and 25.84% respectively. It shows that the FS method gives better predictive performance than the FC method for the estimation of GHI. [ABSTRACT FROM AUTHOR]

• Integrated CPFEM-ML method is developed for defect induced fatigue life prediction. • Data augmentation with FIPs is used to predict fatigue life with different R ratios. • CNN achieves high accuracy in fatigue life prediction for microstructural features. Fatigue life prediction based on the traditional empirical formulation in large fatigue regimes is limited and prone to over-fitting. In this regard, we propose an integrated approach for fatigue life prediction of lamellar titanium alloys incorporating crystal plasticity simulations and a variety of machine learning methods from a microstructural perspective. Firstly, various defect shapes and distinct lamellar distributions are employed to characterize fatigue performance under different microstructural features. By utilizing fatigue indicator parameters (FIPs) to get an expanded dataset, this work employs algorithms including artificial neural networks (ANN), random forests regression (RFR), support vector regression (SVR), and particularly convolutional neural networks (CNN) to predict fatigue life. The results indicate that CNN outperforms the previously mentioned algorithms in prediction accuracy, due to its ability to directly extract microstructural features from images. Our proposed approach provides a microstructural perspective for fatigue life prediction of lamellar titanium alloys, broadening the way for life assessment of materials with diverse microstructures. [ABSTRACT FROM AUTHOR]

• The correlation between TBM data and rock mass classification was investigated. • Input parameters for a rock mass classification prediction model were identified. • The dynamic relationships between rock masses and parameters were quantified. • Rock mass classifications were predicted using the nomogram classification model. • The accuracy and practical convenience of this prediction method are demonstrated. With the widespread application of tunnel boring machines (TBM) in underground tunnel construction, the accurate and real-time identification of geological indicators has become a key factor for ensuring construction safety. In this study, we introduced the nomogram method into the field of civil engineering for the first time to predict surrounding rock grades. Against the backdrop of a major water conservancy project in China, an in-depth analysis of construction data from the ascending phase was conducted to establish the relationship between surrounding rock grade predictions and key construction parameters. Compared with traditional models, the nomogram method has shown advantages in terms of prediction accuracy. Furthermore, by providing a personalized scoring mechanism for specific construction data, the nomogram method not only achieves the real-time prediction of rock grades but also clearly reveals the dynamic relationships between input parameters and rock grades, clarifying the entire prediction process. We identified the main factors related to rock grade prediction among TBM construction parameters, including the slurry conveyor belt speed, jacking pressure, cutterhead torque, cutterhead thrust, thrust pressure, and actual rotation speed. The training process revealed that the nomogram prediction model achieved optimal performance with seven input parameters, balancing accuracy and training time. According to the prediction results, the nomogram model's area under the curve index is 0.92, providing greater accuracy than the traditional random forest model (0.89). Therefore, this study enhances the transparency and interpretability of model predictions by providing personalized assessments of construction parameters for predicting surrounding rock grades, overcoming the black-box issue of existing prediction methods and offering a novel and effective tool for rock grade prediction. [ABSTRACT FROM AUTHOR]

Improvement of data-driven techniques, specifically machine learning (ML), in material science turned it into a powerful tool for predicting materials behavior. Accordingly, this study provides a ML prediction of empirical creep lifetimes of 9Cr-1Mo ex-service heater tubes that have been used in industry for up to 47 years. Data from over 90,000 h of stress rupture tests shows that the service parameters influence creep lifetime similar to mechanical properties. Employing six different ML algorithms, viz., K Nearest Neighbors (KNN), Support Vector Regressor (SVR), Random Forest (RF), Gradient Boosting (GB), Gaussian Process (GP), and Multi-Layer Perceptron (MLP) demonstrated that the GP and MLP methods performed significantly better in predicting the creep lifetimes rather than other algorithms. Finally, a validation set involving 12 samples was conducted, and the GP algorithm showed better agreement with experimental values than other ML and Larson-Miller Parameter approaches, illustrating the capability of this model to predict creep lifetimes. [Display omitted] [ABSTRACT FROM AUTHOR]

• This is the first research of calibrateing quantile regression forests with conformal predictive systems for better performance. • Out-of-bag predictions of quantile regression forests are employed to make full use of data and improve computational efficiency. • With analysis of quantile regression forests, our proposed weighed approach performs best in the comparison experiments. Quantile regression forests (QRF) is a generalization of random forests for quantile regression, which can also output probabilistic prediction for regression problems. QRF delivers a nonlinear and nonparametric way of probabilistic prediction and is widely used in many applications. To improve the prediction quality of QRF, this paper builds conformal predictive systems (CPSs) on top of QRF to calibrate the probabilistic prediction, which is the first attempt of combining CPSs with QRF to our best knowledge. Conformal predictive systems aim to produce valid probabilistic predictions in machine learning community, whose predictive distributions are compatible with realizations. Three algorithms are proposed in this paper. One is a fast approximation algorithm to split conformal predictive system. Building on that, the other two algorithms are proposed based on the out-of-bag samples and the weighted approach of QRF. Experiments with 20 public data sets were conducted and the results demonstrated that our algorithms are empirically valid and compared favourably with the original QRF. [ABSTRACT FROM AUTHOR]

Predicting vessel trajectories is crucial for enhancing situational awareness and preventing collisions at sea. However, achieving accurate and efficient predictions is challenging due to the heterogeneity in vessel movement patterns and changes in vessel mobility modes during voyages. To address this, we propose a new approach that uses historical AIS data to cluster route patterns for each vessel type, thereby improving prediction accuracy. By training machine learning algorithms to focus only on similar vessel types, this approach can better predict individual vessel mobility patterns. This approach offers computational advantages by using a relatively small set of trajectories from the nearest cluster of a selected vessel. Both spatial and course attributes are considered to determine the nearest cluster, while engineered features capture changes in vessel mobility modes. Using an AIS dataset from UTM Zone 10N (US West Coast), we achieved distance errors of 370 m , 742 m , and 1. 2 k m for horizons 10, 20, and 30 min, respectively, using the Random Forest algorithm for short-term trajectory prediction (≤ 30 min) with the last 1-hour trajectory of selected vessels as input. • Proposed framework for predicting the short-term future trajectory of vessels. • Enhanced prediction accuracy by addressing heterogeneity in mobility patterns. • Engineered features are incorporated to consider changes in mobility modes. • Spatial and course attributes are used for trajectory classification. • Evaluation shows high prediction accuracy with low computational cost. [ABSTRACT FROM AUTHOR]

Data-driven model development brings new approaches to solve conventional civil engineering problems, which are usually considered and answered by experimental, analytical, and numerical methods. This study aims to develop an ensemble learning model (i.e., XGBoost: eXtreme Gradient Boosting) to predict ultimate moment of reinforced concrete (RC) members strengthened by a newly developed concrete technology – ultrahigh performance concrete (UHPC). The study considered two scenarios, incorporating eighteen and seventeen features, with one feature modification involving the transformation of width and height to the cross-sectional area of RC members. Incorporating three substrate damage levels, two substrate surface treatments preceding UHPC strengthening, and five UHPC layouts as model features, the hyperparameters of the XGBoost model are fined-tuned through random search and k-fold cross-validation techniques. Model performance is evaluated using metrics of mean-absolute error (MAE), mean-square error (MSE), root-mean-square error (RMSE), and coefficient of determination (R 2). The XGBoost model, trained on a dataset of170 instances, achieves an R 2 accuracy of 92.5% in the training set and 81.9% in the unseen testing set. Based on the feature significance score, the top four most influential features were reinforcement ratio in UHPC layer (ρ u) , longitudinal reinforcement ratio in RC member (ρ sl ) , UHPC thickness (hu), and RC member height (h), collectively contributing more than half of the model's predictive power. Feature engineering does not yield significant benefits. In comparison to other tree-based ensemble learning models (Random Forest, AdaBoost, Gradient Boosting, and Bagging), the developed XGBoost model, demonstrates superior overall prediction performance. This research has practical implications for the design and analysis of UHPC strengthened RC members. The model's accuracy is expected to improve with additional data collection to address imbalances in feature distributions within the current limited dataset. • An ensemble model-XGBoost is developed to predict ultimate moment of RC members strengthened by UHPC for the first time. • Eighteen features were considered in the machine learning models. • XGBoost model shows better prediction than Random Forest, AdaBoost, Gradient Boosting and Bagging. • Reinforcement ratios in UHPC layer and RC member, UHPC thickness, and RC thickness are top four influential features. [ABSTRACT FROM AUTHOR]

This study presents a comprehensive and systematic review of the machine learning (ML) techniques used to predict employee turnover in the past decade. A total of 52 relevant peer-reviewed studies published between 2012 and April 2023 were selected. The results indicate that over 20 ML techniques have been used to predict employee turnover in various institutions. In addition, this comprehensive review demonstrates that most machine learning approaches used to predict employee turnover were based on supervised learning, with 96% of the articles (50 out of 52) in this category, random forest technique. Furthermore, the review reveals that the most critical factors for predicting employee turnover have been salary and overtime. This study makes a valuable contribution to the field by offering a systematic analysis of the ML algorithms used for predicting employee turnover, in addition to providing an overview of the most significant works in this field produced in the past decade. This study offers important reference regarding the essential ML approaches used in employee turnover prediction and provides future directions for researchers and industries. [ABSTRACT FROM AUTHOR]

The length-scale effect on the hardness of Al/Si 3 N 4 multilayered composites were investigated via combined experimental and machine learning approaches. It was found that the hardness decreases with increase of the individual layer thickness, in agreement with the classic Hall-Petch relation. However, the hardness decrease became moderate when the individual layer thickness was less than 80 nm. The individual critical layer thickness of metallic/ceramic multilayered composites for the change in deformation mechanism was obtained by machine learning methods. The prediction results indicated that the individual critical transition layer thickness was mainly determined by the metal layer, and the corresponding critical transition hardness was positively correlated with the elastic modulus of the ceramic layer. Importantly, it was confirmed that the hardness of other metallic/ceramic multilayered composite systems could be predicted by the machine learning approach, especially the optimized random forest regression (RFR) model. [Display omitted] [ABSTRACT FROM AUTHOR]

The prediction of thermal conductivity of aligned arrangement ellipsoidal particles reinforced composites is realized by a data-driven method based on machine learning models. The models of random forest regression, polynomial regression, gradient boosting regression, and multi-layer perceptron regression are compared and analyzed. Furthermore, the effects of the thermal conductivity ratio of matrix and particle, volume fraction, aspect ratio, orientation angle of particles, and interfacial thermal resistance on the thermal conductivity of composites were analyzed by machine learning model. The results show that the data-driven model based on random forest regression has the highest accuracy in predicting the thermal conductivity of ellipsoidal particle reinforced composites, with R 2 = 0.994 and MAPE = 1.82%. However, the model has errors in the details of predicting the trend of thermal conductivity change. Therefore, the data-driven model is used to calibrate the parameters of the theoretical model. The modified theoretical model combines the advantages of high numerical accuracy of the machine learning model with the advantages of accurate prediction of the thermal conductivity trend of the theoretical model. • A data-driven model was constructed to accurately predict the thermal conductivity of composites. • The effects of multiple parameters on thermal conductivity were quantitatively evaluated using the random forest model. • For the first time, the data-driven model is used to construct the theoretical model, improving its accuracy. [ABSTRACT FROM AUTHOR]

[Display omitted] • HHV of fuel is the most critical parameter for adopting waste-to-energy strategies. • Prediction of HHV of fuel using machine learning models and interpretability analysis. • Comparison of six different ML models (MLR, ETR, RFR, SVR, DTR, XGB) • ETR is the best-performing model according to R2, RMSE & MAPE scores. • Local explanation for predictions and interaction effect of variables using SHAP. This study presents a novel approach using machine learning techniques to estimate waste materials' higher heating value (HHV), which plays a crucial role in waste-to-energy generation efficiency. The study utilizes a dataset comprising ultimate and proximate analysis of 16 different waste types. It employs six machine learning models: Extra Trees, Random Forest, Support Vector, Decision Tree, Extreme Gradient Boosting, and Multivariate Linear Regressors. The investigation explores the relationships between the features and outcomes through Spearman correlation, feature importance analysis, SHAP dependence, and decision plots, providing the interpretability of the model's predictions. The models are fine-tuned with hyperparameters for six feature sets, enabling researchers to anticipate HHV based on their specific input. The results demonstrate high accuracy in predicting HHV, with R2 ranging from 0.83 to 0.98, RMSE from 2.25 to 0.79, and MAPE from 6.01 to 0.92%. The study further reveals that higher carbon and hydrogen content increases HHV, while higher oxygen and ash content results in decreased HHV. Notably, Carbon, Ash content and Hydrogen content are the significant features with mean absolute SHAP values of 2.17, 0.65, and 0.37, respectively. The proposed alternative prediction method has practical implications for waste-to-energy generation research and practice, facilitating informed decision-making in this field. [ABSTRACT FROM AUTHOR]

[Display omitted] • Dataset containing 3830 particles from lightweight packaging waste created. • Particle masses vary significantly between material and particle size classes. • Volume-equivalent diameter correlates best with particle mass. • Best machine learning model outperforms state-of-the-art by 43% higher R 2-score. • Random forest regressor achieves best mass predictions with R 2 = 0.763 ± 0.091. Sensor-based material flow characterization (SBMC) promises to improve the performance of future-generation sorting plants by enabling new applications like automatic quality monitoring or process control. Prerequisite for this is the derivation of mass-based material flow characteristics from pixel-based sensor data, which requires known individual particle masses. Since particle masses cannot be measured inline, the prediction of particle masses of lightweight packaging (LWP) waste using machine learning (ML) algorithms is investigated. Five LWP material classes were sampled, preprocessed, and scanned on a custom-made test rig, resulting in a dataset containing 3D laser triangulation (3DLT) images, RGB images, and corresponding masses of n = 3,830 particles. Based on 66 extracted shape measurements, six ML models were trained for particle mass prediction (PMP). Their performance was compared with two state-of-the-art reference models using (i) material-specific mean particle masses and (ii) grammages. Obtained particle masses showed a high variation and significant differences between material classes and particle size classes. After feature selection, both reference models achieving R 2-scores of (i) 0.422 ± 0.121 and (ii) 0.533 ± 0.224 were outperformed by all investigated ML models. A random forest regressor with an R 2-score of 0.763 ± 0.091 and a normalized mean absolute error of 0.243 ± 0.050 achieved the most accurate PMP. In contrast to studies on primary raw materials, PMP of LWP waste is challenging due to influences of packaging design and post-consumer disposal behavior. ML algorithms are a promising approach for PMP that outperform state-of-the-art methods by 43% higher R 2-scores. [ABSTRACT FROM AUTHOR]

• Performance assessment and production estimation of grid-connected PV station. • Strong relationship between weather variables and the output of the PV station. • Strong interdependence of the meteorological parameters and performance parameters. • The distinction between the different variable correlation results is a hard task. • Feature selection and PCA analysis were used as dimensional reduction techniques. • Dimensionality reduced data was used for training random forest models. Meteorological variables have an important effect on the performance of a grid-connected photovoltaic station, in this paper, the impact of meteorological variables on the 6 mWp grid-connected photovoltaic station in the desert of Adrar has been explored through performance assessment and output power forecasting. The impact of the meteorological variables on performance parameters has been investigated by performing interdependence and correlation studies. A clear interdependence between some variables has been observed, but the complete separation between each variable correlation effect has proven to be a difficult task. A combination of random forest method and pre-processing techniques namely feature selection and Principal component analysis has been developed in order to predict power production using meteorological variables as inputs. The forecasting models have been evaluated in terms of computation time, accuracy, and several statistical indicators. [ABSTRACT FROM AUTHOR]

The scale-up of nanocellulose production, a still challenging task, could benefit greatly from machine learning. In this paper, random forest regressors were applied to predict two important parameters for nanocellulose processing: nanofibrillation yield and aspect ratio. The dataset comprised diverse pulp sources, ranging from non-woody and woody feedstock, and two different types of pulp pre-treatments, namely mechanical refining and enzymatic hydrolysis. The obtained models were evaluated in terms of number of features and optimized with regard to some important hyperparameters, using the Bayesian optimization function from the SkOpt library. Considering the heterogeneity of the data, the prediction of both aspect ratio and yield of nanofibrillation was deemed satisfactory. For the former, R² scores of 0.986, 0.902 and 0.932 were obtained for the training, validation and testing subsets, respectively. Regarding the latter, R² scores of 0.994, 0.947 and 0.877 were achieved for the same types of subsets. Other important remarks are that enzyme dosage was the most important feature for predicting aspect ratio, while it did not play an important role for yield modelling. For this output, the transmittance at 600 nm accounted for almost 90% of the total feature importance. [Display omitted] • Aspect ratio and nanofibrillation yield were predicted by random forest regressors. • A 140-point dataset was gathered from experiments from different woods and crops. • Both independent variables and directly measurable responses constituted the inputs. • Bayesian optimization helped attain R2 of 0.90–0.95 for the validation subsets. • The kind of feedstock had lower feature importance than the type of pretreatment. [ABSTRACT FROM AUTHOR]

Accurate load forecasting is important to mitigate the negative impact of Electric vehicle integration into the existing grid. Previous studies mostly focus on individual or aggregated levels without specifying the impact of accuracy due to the selection of different spatial levels and lack the integration of uncertainty estimation in the forecasting models. To address these issues, this study compares the predictive performance of a Random Forest and Artificial Neural Networks at different spatial levels with 15-min resolution data across case studies (i) with 2 Electric Vehicles charging poles and 3 users, (ii) with 75 charging poles, 8 charging rails and 70 users. The outcome shows that forecasting the Electric Vehicle load of smaller case studies will require the presence or calendar information of users. Whereas in case studies with more than 10 charging piles, the features "previous week's power", "hour of the day" and the "number of connections" can achieve similar results. The results also showed that the aggregated forecasting was more accurate than individual charging piles. Moreover, the uncertainty plot generated for a 90% prediction interval showed that the uncertainty estimates were more reliable for the case study with large numbers of Electric Vehicles. • Framework for Electric Vehicle load forecasting with spatial considerations. • Integration of uncertainty estimation using quantile regression. • Comparative study between Random Forest and Artificial Neural Networks models. • Aggregated Electric Vehicle load forecasting outperforms individual charging piles. [ABSTRACT FROM AUTHOR]

• An integrated approach that combines machine learning with Bayesian spatial Poisson model is proposed. • The study predicts large-scale real-time occurrence and frequency of traffic conflicts with different severities. • The spatial correlation across the study intersections is accounted for. • Several novel spatiotemporal traffic state measures are developed. The use of traffic conflicts in road safety evaluation is gaining considerable popularity as it plays a vital role in developing a proactive safety management strategy and allowing for real-time safety analysis. This study proposes an integrated approach that combines a machine learning (ML) algorithm and a Bayesian spatial Poisson (BSP) model to conduct large-scale real-time traffic conflict prediction by considering traffic states as the explanatory variables. Traffic conflicts are measured by two indicators, the Time to Collision (TTC) and the Post-Encroachment Time (PET). Based on both TTC and PET, traffic conflict severity is classified into five categories. For each conflict severity category, a binary variable (conflict occurrence) and a count variable (conflict frequency) are developed, respectively. In addition to conflict variables, traffic state parameters are extracted from a large-scale high-resolution trajectory dataset. The traffic parameters include volume, density, speed, and the corresponding space-based and space–time-based measures within a 30-second interval. Eight ML-based classifiers are applied to predict conflict occurrence, and the best classifier is selected. A binary logistic regression is developed to explore the potential linkages between traffic states and conflict occurrence. As well, a resampling technique Borderline-SMOTE is used to mitigate the sparsity caused by the predefined short interval. The BSP model is utilized to predict the specific number of conflicts. Further, the BSP model can also explain the relationship between traffic states and conflict frequency, and thus the significant influencing traffic states are identified. The results show that random forest outperforms the other MLs in terms of conflict occurrence prediction accuracy. Further, the proposed integrated approach achieves a high performance of conflict frequency prediction with RMSE values of 0.1384 ∼ 0.1699, MAPE values of 9.25% ∼ 36.99%, and MAE values of 0.0087 ∼ 0.6398. The finding emphasizes the need for separately predicting the occurrence and frequency of conflicts with different severities. [ABSTRACT FROM AUTHOR]

This paper presents three ensemble learning models for short term load forecasting. Machine learning has evolved quickly in recent years, leading to novel and advanced models that are improving the forecasting results in multiple fields. However, in highly dynamic fields such as power and energy systems, dealing with the fast acquisition of large amounts of data from multiple data sources and taking advantage from the correlation between the multiple available variables is a challenging task, for which current models are not prepared. Ensemble learning is bringing promising results in this sense, as, by combining the results and use of multiple learners, is able to find new ways for current learning models to be used and optimized. In this paper three ensemble learning models are developed and the respective results compared: gradient boosted regression trees, random forests and an adaptation of Adaboost. Results for electricity consumption forecasting in hour-ahead are presented using a case-study based on real data from an office building. Results show that the adapted Adaboost model outperforms the reference models for hour-ahead load forecasting. [ABSTRACT FROM AUTHOR]

Background: Depression and post-traumatic stress disorder (PTSD) are among the most common psychiatric disorders observed in children and adolescents exposed to sexual abuse.Objective: The present study aimed to investigate the effects of many factors such as the characteristics of a child, abuse, and the abuser, family type of the child, and the role of social support in the development of psychiatric disorders using machine learning techniques.Participants and Settings: The records of 482 children and adolescents who were determined to have been sexually abused were examined to predict the development of depression and PTSD.Methods: Each child was evaluated by a child and adolescent psychiatrist in the psychiatric aspect according to the DSM-V. Through the data of both groups, a predictive model was established based on a random forest classifier.Results: The mean values and standard deviation of the 10-k cross-validated results were obtained as accuracy: 0.82% (+/- 0.19%), F1: 0.81% (+/- 0.19%), precision: 0.81% (+/- 0.19%), recall: 0.80% (+/- 0.19%) for children with depression; and accuracy: 0.72% (+/- 0.12%), F1: 0.71% (+/- 0.12%), precision: 0.72% (+/- 0.12%), recall: 0.71% (+/- 0.12%) for children with PTSD, respectively. ROC curves were drawn for both, and the AUC results were obtained as 0.88 for major depressive disorder and 0.76 for PTSD.Conclusions: Machine learning techniques are powerful methods that can be used to predict disorders that may develop after sexual abuse. The results should be supported by studies with larger samples, which are repeated and applied to other risk groups. [ABSTRACT FROM AUTHOR]

The objectives of the study were to use a heat stress scoring system to evaluate the severity of heat stress on dairy cows using different heat abatement techniques. The scoring system ranged from 1 to 4, where 1 = no heat stress; 2 = mild heat stress; 3 = severe heat stress; and 4 = moribund. The accuracy of the scoring system was then predicted using 3 machine learning techniques: logistic regression, Gaussian naïve Bayes, and random forest. To predict the accuracy of the scoring system, these techniques used factors including temperature-humidity index, respiration rate, lying time, lying bouts, total steps, drooling, open-mouth breathing, panting, location in shade or sprinklers, somatic cell score, reticulorumen temperature, hygiene body condition score, milk yield, and milk fat and protein percent. Three different treatments, namely, portable shade structure, portable polyvinyl chloride pipe sprinkler system, or control with no heat abatement, were considered, where each treatment was replicated 3 times with 3 second-trimester lactating cows. Results indicate that random forest outperformed the other 2 methods, with respect to both accuracy and precision, in predicting the sprinkler group's score. Both logistic regression and random forest were consistent in predicting scores for control, shade, and combined groups. The mean probability of predicting non-heat-stressed cows was highest for cows in the sprinkler group. Finally, the logistic regression method worked best for predicting heat-stressed cows in control, shade, and combined. The insights gained from these results could aid dairy producers to detect heat stress before it becomes severe, which could decrease the negative effects of heat stress, such as milk loss. [ABSTRACT FROM AUTHOR]

Suicide is a serious social and public health problem. Social stigma and prejudice reduce the accessibility of mental health care services for high-risk groups, resulting in them not receiving interventions and committing suicide. A suicide prediction model is necessary to identify high-risk groups in the general population. We used national medical check-up data from 2009 to 2015 in Korea. The latest medical check-up data for each subject was set as an index point. Analysis was undertaken for an overall follow-up period (index point to the final tracking period) as well as for a one-year follow-up period. The training set was cross-validated fivefold. The predictive model was trained using a random forest algorithm, and its performance was measured using a separate test set not included in the training. The analysis covered 372,813 individuals, with an average (SD) overall follow-up duration of 1.52 (1.52) years. When we predicted suicide during the overall follow-up period, the area under the receiver operating characteristic curve (AUC) was 0.849, sensitivity was 0.817, and specificity was 0.754. The performance of the predicted suicide risk model for one year from the index point was AUC 0.818, sensitivity 0.788, and specificity 0.657. This is probably the first suicide predictive model using machine learning based on medical check-up data from the general population. It could be used to screen high-risk suicidal groups from the population through routine medical check-ups. Future studies may test preventive interventions such as exercise and alcohol in these high-risk groups. [ABSTRACT FROM AUTHOR]

Since the cost and time required to finetune parameters in traditional short-term traffic prediction models such as traffic simulators are very high, the prediction models have been developed mainly for managing recurrent congestion, rather than non-recurrent congestion caused, for example, by disaster. Machine learning models are promising candidates for traffic prediction during non-recurrent congestion due to their ability to tune parameters without a-priori knowledge, while their applicability to non-recurrent conditions has rarely been explored. To fill in this gap, this study conducts an exploratory analysis on the applicability of various machine learning models during a transportation network disruption with particular focuses on their ability to predict traffic states and the interpretability of the results. The analysis is conducted by using data obtained during the massive transport network disruption which occurred in Hiroshima in July 2018 due to heavy rain and subsequent landslides. The models tested include random forest, support vector machine, XGBoost, shallow feed-forward neural network, and deep feed-forward neural network. The results indicate that random forest and XGBoost methods produced the best results in terms of prediction accuracy. On the other hand, deep neural network models produce better results in terms of the interpretability of the results, i.e., the results can be logically explained from the perspective of existing traffic flow theory. These findings indicate that the model which produces the best prediction accuracy is not always the best for practical use since it does not mimic the mechanisms of congestion occurrence. • Apply machine learning models to non-recurrent congestion caused by disaster. • Explore their ability to (1) predict traffic states and (2) interpret the results. • XGBoost method produced the best results in terms of prediction accuracy. • Deep neural network models are better in terms of the interpretability. • Confirm the best prediction accuracy is not always the best for practical use. [ABSTRACT FROM AUTHOR]

The stopping power of a material upon interaction with an energetic ion is the key measure of how far that ion will travel. The implications of accurate particle range calculations are tremendous, affecting every single application in which particle radiation is involved, from nuclear power to medicine. An approach is presented which attempts to overcome current shortcomings in the theoretical understanding of stopping power, as well as the methods used to interpret and exploit measured data. This is a considerable challenge, however the use of a novel machine learning methodology is shown to hold great promise in this endeavour: the ultimate aim being the ability to correctly predict the stopping value for any energy, ion and target combination, having no pre-existing experimental data. A random forest regression algorithm is trained using over 34,000 experimental measurements, representing stopping power values for 522 ion-target combinations across the energy range 10 - 3 to 10 2 MeV/amu, and ion and target atomic masses 1 to >240. Evaluation is carried out using several fundamental error metrics, over the whole dataset as well as for individual combinations, to provide the most comprehensive understanding of performance when tested under strict cross-validation criteria. The resulting model is shown to yield predicted stopping power curves corresponding closely to those of the true experimental values, with an ability to generalise across target elements, compounds, mixtures, alloys and polymers, irrespective of phase, and for a wide range of ion masses. [ABSTRACT FROM AUTHOR]

Quantitative Structure-Property Relationship (QSPR) classification models for Minimum Ignition Energy (MIE) classification of 65 combustible dusts have been developed using machine learning algorithms Random Forests (RF) and Decision Trees (DT). Combustible dusts were categorized based on their MIE values: dusts having MIE > 10 mJ and dusts having MIE < 10 mJ. In this work, an important macroscopic property, dust median diameter (d 50) was also used as one of the parameters along with the molecular descriptors for model development. The machine learning Random Forest (RF) algorithm was used for identification of the 13 molecular descriptors having maximum effect on MIE prediction accuracy. Thereafter, these descriptors were used as input parameters to develop binary classification Random Forest (RF) and Decision Tree (DT) models. The RF and DT algorithms displayed good MIE category predictability validated through the test set Receiver Operating Characteristic- Area Under Curve (ROC-AUC) of 0.95 for both the models. Thus, this work has displayed that Random Forests can be used as an important tool for descriptor reduction in QSPR studies and can prove to be a valuable resource for MIE category prediction. Unlabelled Image • QSPR models for MIE prediction of combustible dusts were developed. • Combustible dusts were categorized into two ignition energy categories. • Classification models developed using Random Forests and Decision Trees. • Dust median diameter (d50) used as one of the input model parameters. • The RF and DT algorithms demonstrate good MIE predictability. [ABSTRACT FROM AUTHOR]

• Use feature selection to extract the most discriminating features. • Improve accuracy and reduce bias using an ensemble technique. • A One-vs-Rest Consensus Learning (OvRCL) technique is proposed. Research in crash severity prediction is necessary to allow safety planners to take precautionary measures and enable first aiders to remain prepared for assisting the injured. Existing literature in the field of crash severity prediction is mostly focused on generating the attributes for predicting the severity. However, in reality, not all features are discriminating, and certain classes are challenging to detect even employing the entire feature set. Although to tackle these problems several techniques are developed in the Machine Learning (ML) literature. But their application to crash severity prediction and an optimal strategy for the best combination of features or classifiers for achieving high accuracy is a less studied area. To address these problems, this work first provides a comparison of widely used classifiers for predicting crash severity; and secondly, by combining class-wise majority voting with One-vs-Rest (OvR) approach, a novel classification framework named, OvR consensus learning (OvRCL) is proposed. The proposed method avail a feature selection technique, Mutual information (MI), to acquire the most relevant feature set regarding the output class (i.e. severity). Moreover, to differentiate each class in the multi-class data, OvRCL iteratively runs ML algorithms as binary classifiers in an ensemble framework to significantly ameliorate classification performance. In our experiments, we use four classifiers, namely, the k -Nearest Neighbors (k -NN), Support Vector Machines (SVM), Random Forest (RF), and Bagging classifier, to get the consensus. Analysis was done using a real crash dataset obtained from an open data source of Leeds city council. A four-year crash data (2015–2018) is used for training and the OvRCL is tested on the 2019 data. Moreover, to validate the performance of OvRCL, this study also utilizes two more datasets with high-class imbalance. In contrast to conventional ML algorithms, our experiments depict that the OvRCL is a potent method for forecasting crash severity levels on the data under test. [ABSTRACT FROM AUTHOR]

Accurate predictions of source resolved atmospheric PM 2.5 concentrations at high resolutions using chemical transport models (CTMs) require expensive CTM simulations and development of high-resolution emissions inventories. We use multiple machine learning (ML) approaches to downscale coarse-resolution (36 × 36 km2) CTM predictions to 1 × 1 km2 spatial resolution. ML predictions include concentrations of the major chemical components of PM 2.5 and the contributions of its major emissions sources. Inputs for the ML models include 36 × 36 km2 source resolved CTM predicted concentrations of all PM 2.5 components, meteorological data, and several land-use (LU) variables. The output of our ML models is the 1 × 1 km2 source-resolved concentrations of all major PM 2.5 components in southwestern Pennsylvania (5184 km2 domain) during February and July 2017. Models were trained and validated using 1 × 1 km2 resolution source- and species-resolved CTM predictions of PM 2.5 from recent complementary studies. The best overall performance was found using a random forest (RF) model, where species and source resolved PM 2.5 concentrations were reproduced with low normalized mean bias (|NMB| < 0.01). The downscaling model captures the spatial distribution of PM 2.5 both by component and source, with some discrepancies when predicting the plumes of large point sources that have long-range impacts. In a test of generalizability to unknown domains, the model differentiates well between areas that are primarily urban, rural, or industrial but faces challenges with the reproduction of the effects of large point sources of PM 2.5 when entire quadrants are removed from the training data. The results represent a proof of concept for downscaling low-resolution CTM predictions using native high-resolution CTM predictions in training. • New ML method for downscaling chemical transport model predictions. • Provides particle mass concentration, composition, and source contributions. • Does not require high-resolution emission inventories. • Uses as inputs available land use regression (LUR) variables. • Encouraging results in its first application in Pittsburgh, US. [ABSTRACT FROM AUTHOR]

When a forecast of the total value over several time periods ahead is required, forecasters are presented with two temporal aggregation (TA) approaches to produce required forecasts: i) aggregated forecast (AF) or ii) aggregate data using non-overlapping temporal aggregation (AD). Often, the recommendation is to aggregate data to a frequency relevant to the decision the eventual forecast will support and then produce the forecast. However, this might not be always the best choice and we argue that both AF and AD approaches may outperform each other in different situations. Moreover, there is a lack of evidence on what indicators may determine the superiority of each approach. We design and execute an empirical experiment framework to first explore the performance of these approaches using monthly time series of M4 competition dataset. We further turn the problem into a classification supervised learning by constructing a database consisting of features of each time series as predictor and model class labelled as AF/AD as response/outcome. We then build machine learning algorithms to investigate the association between time series features and the performance of AF and AD. Our findings suggest that both AF and AD approaches may not consistently generate accurate results for every individual series. AF is shown to be significantly better than AD for the monthly M4 time series, especially for longer horizons. We build several machine learning approaches using a set of extracted time series features as input to predict accurately whether AD or AF should be used. We find out that Random Forest (RF) is the most accurate approach in correctly classifying the outcome assessed both by statistical measures such as misclassification error, F-statistics, area under the curve, and a utility measure. The RF approach reveals that curvature, nonlinearity, seas_pacf, unitroot_pp, mean, ARCHM.LM, Coefficient of Variation, stability, linearity, and max_level_shif are among the most important features in driving the predictions of the model. Our findings indicate that the strength of trend, ARCH.LM, hurst, autocorrelation lag 1, unitroot_pp, and seas_pacf may favour AF approach, while lumpiness, entropy, nonlinearity, curvature, and strength of seasonality may increase the chance of AD performing better. We conclude the study by summarising the findings and present an agenda for further research. [ABSTRACT FROM AUTHOR]

This work contributes to this ongoing debate on the nature and the characteristics of propagation channels of crash events in international stock markets. Specifically, we investigate transmission mechanisms across stock markets along with effects from bond and currency markets. Our approach comprises a solid forecasting mechanism of the probability of a stock market crash event in various time frames. The developed approach combines different machine learning algorithms which are presented with daily stock, bond and currency data from 39 countries that cover a large spectrum of economies. Specifically, we leverage the merits of a series of techniques including Classification Trees, Support Vector Machines, Random Forests, Neural Networks, Extreme Gradient Boosting, and Deep Neural Networks. To the best of our knowledge, this is the first time that Deep Learning and Boosting approaches are considered in the literature as a means of predicting stock market crisis episodes. The independent variables included in our data contain information regarding both the two fundamental linkage channels through which financial contagion can be initiated: returns and volatility. We apply a suite of machine learning algorithms for selecting the most relevant variables out of a large set of proposed ones. Finally, we employ bootstrap sampling for adjusting the imbalanced nature of the available fitting dataset. Our experimental results provide strong evidence that stock market crises tend to exhibit persistence. We also find significant evidence of interdependence and cross-contagion effects among stock, bond and currency markets. Finally, we show that the use of Deep Neural Networks significantly increases the classification accuracy, while offering a robust way to create a global systemic early warning tool that is more efficient and risk-sensitive than the currently established ones. Thus, central banks may use these tools to early adjust their monetary policy, so as to ensure financial stability. [ABSTRACT FROM AUTHOR]

Highlights • Machine learning models are compared for simulation and forecasting of streamflows. • Support Vector Regression, Artificial Neural Networks and Random Forest are used. • A base flow separation method is coupled with machine learning models. • North Fork, Chehalis, Carson and Sacramento rivers in the United States are analysed. • Base flow separation method improved streamflow simulation to a certain degree. Abstract Efficient simulation of rainfall-runoff relationships is one of the most complex problems owing to the high number of interrelated hydrological processes. It is well-known that machine learning models could fail in simulating streamflows from only meteorological variables in the absence of antecedent streamflow values. The main reason for this could be low and lagged relationships between streamflow and meteorological variables. To overcome this inefficiency, for the first time, we developed a simulation framework by coupling a base flow separation method to three machine learning methods. It was demonstrated that separating streamflow into different components such as base flow and surface flow can be useful for improving simulation and forecasting capabilities of machine learning models. We simulated streamflow in four rivers in the United States with Support Vector Regression (SVR), Artificial Neural Networks (ANNs) and Random Forest (RF) as a function of precipitation (P), temperature (T) and potential evapotranspiration (PET). We concluded that the base flow separation method improved the simulation performances of the machine learning models to a certain degree. Apart from the simulation scheme, we also employed a forecasting scheme by using the antecedent observed discharge values in addition to P, T, and PET. We discussed performances of models in simulation and forecasting of streamflow regarding model types, input structures and catchment dynamics in detail. [ABSTRACT FROM AUTHOR]

The main objective of the present paper is to demonstrate the feasibility of machine-learning-based surrogate models for predicting low-velocity impact behaviour considering void content and location generated during the resin transfer moulding process. Generating reliable experimental datasets for training those models is almost impossible, therefore an adapted finite element model was implemented providing reliable results to generate the synthetic datasets. The optimum hyperparameter combination for training the Random Forest model was found based on the grid search technique. The accuracy of the classification, single-output regression and multi-output regression models was sufficient. It was concluded that the multi-output regression model, which predicts the force-time, displacement-time, and energy-time curves, provides the best information, and is sufficiently accurate (R2 > 0.995) and fast (5 s per sample) as an online structural performance monitoring tool. [ABSTRACT FROM AUTHOR]

Fuel consumption influences both the economic and environmental perspectives of shipping. With the help of machine learning, meaningful knowledge and complex relationships can be extracted from high-dimensional historical data. In this study, machine learning models were developed to predict the fuel consumption of harbour vessels with ship-related and meteorological factors. The superiority of machine learning models over statistical linear regression model (Ridge regression) has been proved. This study further investigated whether the use of meteorological factors enhances the prediction of fuel consumption. A case study on the prediction of tugboat fuel consumption was conducted. The Random Forest model outperformed the other models. Comparative experiments showed that meteorological factors collectively add value to the fuel consumption prediction, which enhances the accuracy from 0.7 to 38.9%. The potential uses of the prediction results are highlighted in terms of both operational management and environmental evaluation aspects. • Machine learning models were developed to predict the fuel consumption of harbour vessels. • The superiority of machine learning models over statistical linear regression model (Ridge regression) has been proved. • The Random Forest model outperformed the other models. • The meteorological factors collectively add value to fuel consumption prediction, which enhances the accuracy from 0.7 to 38.9%. [ABSTRACT FROM AUTHOR]

• ML framework for fatigue life prediction of AM Ti-6Al-4V samples is proposed. • ANN, RFR and SVR models are used for fatigue life prediction. • Spearman's rank correlation test is applied to identify insensitive features. • The LOOCV technique is employed in the optimization of the ML models. In this work, a framework based on the machine learning (ML) approach and Spearman's rank correlation analysis is introduced as an effective instrument to solve the influence of defects detected by micro-computed tomography (μCT) method, and stress amplitude on the fatigue life performance of AM Ti-6Al-4V. Artificial neural network (ANN), random forest regressor (RFR) and support vector regressor (SVR) models are implemented and optimized. The optimization is performed on training set by tuning the hyperparameters and parameters using the leave-one-out cross validation (LOOCV) technique. The results present comparison between predicted and experimental results and validate the proposed framework. [ABSTRACT FROM AUTHOR]

Millimeter-wave communication systems design requires accurate path-loss (PL) prediction, which is critical to determining coverage area and system capacity. In this work, four machine learning (ML) algorithms are proposed for PL prediction in an indoor environment for 5G millimeter-wave frequencies, from 26.5 to 40 GHz. They are artificial neural network (ANN), support vector regression (SVR), random forest (RF), and gradient tree boosting (GTB). We compare their performances, including also the empirical PL models ABG and CIF incorporating the number of crossed walls, as we propose. For the ML models, we present a methodology to select the predictor coalition by examining the marginal performance and the interpretability gains. The methodology allows choosing the most appropriate predictor coalition for building machines while not overlooking the non-linear connections between predictors and the path-loss. The results show that the ML techniques improve the prediction accuracy of empirical models. Among the ML algorithms employed, considering the optimal subset of predictors, the GTB model presents the best prediction and generalization capacities. • Accurate machine learning regression for the path-loss of millimeter waves indoors. • Incorporating the number of crossed walls improves empirical path-loss models. • Methodology for predictor coalition selection considering interpretability. • The methodology examines the marginal performance and the interpretability gains. • The resulting models present feasible accurate path-loss prediction capacity. [ABSTRACT FROM AUTHOR]

• Multi-label classification models to predict pipe failures in water supply networks. • Avoided pipe failures' percentage raises as the time period to predict for grows. • A uninorm is used to integrate the output probabilities of the different years. The unexpected failure of pipes is a problem that is hitting the water networks of many cities around the world. Nowadays, many proposals based on the use of machine learning techniques are emerging to combat this problem. However, most studies focus their efforts on predicting failures in short time periods, usually a year, while longer time period predictions would be more valuable to address strategic decisions. In this study, the use of multi-label classification techniques is proposed to simultaneously predict pipe failures in water supply systems for multiple years. For this purpose, three models (discriminant analysis, logistic regression and random forest) and different prediction time periods (one, two and three years) have been analysed. As multi-label data require specific quality metrics and sampling techniques, part of this work is dedicated to their exploration and discussion. The models are evaluated on a real-world seven-year database, achieving successful results. An insightful analysis of the use of the methodology shows how the percentage of avoided pipe failures increases over time. In fact, it is demonstrated that 30.2%, 51.4% and 54.0% of the pipe failures of three consecutive years are avoided according to data from a real network. [ABSTRACT FROM AUTHOR]

• Multi-label classification models to predict pipe failures in water supply networks. • Avoided pipe failures' percentage raises as the time period to predict for grows. • A uninorm is used to integrate the output probabilities of the different years. The unexpected failure of pipes is a problem that is hitting the water networks of many cities around the world. Nowadays, many proposals based on the use of machine learning techniques are emerging to combat this problem. However, most studies focus their efforts on predicting failures in short time periods, usually a year, while longer time period predictions would be more valuable to address strategic decisions. In this study, the use of multi-label classification techniques is proposed to simultaneously predict pipe failures in water supply systems for multiple years. For this purpose, three models (discriminant analysis, logistic regression and random forest) and different prediction time periods (one, two and three years) have been analysed. As multi-label data require specific quality metrics and sampling techniques, part of this work is dedicated to their exploration and discussion. The models are evaluated on a real-world seven-year database, achieving successful results. An insightful analysis of the use of the methodology shows how the percentage of avoided pipe failures increases over time. In fact, it is demonstrated that 30.2%, 51.4% and 54.0% of the pipe failures of three consecutive years are avoided according to data from a real network. [ABSTRACT FROM AUTHOR]

• Precipitation δ 18O predicted using geostatistical and machine learning algorithms. • Prediction performance evaluated using ordinary and hybrid error metrics. • Machine learning algorithms performed better, particularly using scarcer datasets. • Two spatiotemporally different subcontinental datasets of precipitation δ 18O were used. Stable isotopes of precipitation are important natural tracers in hydrology, ecology, and forensics. The spatially explicit predictions of oxygen and hydrogen isotopes in precipitation are obtained through different interpolation techniques. In the present study we aim to examine the performance of various interpolation techniques when predicting the spatial distribution of precipitation stable isotopes. The efficiency of combined geostatistical tools (i.e. regression kriging; RK) and various machine learning methods (including regression enhanced random forest methods: MRRF, RERF) are compared in interpolating the spatial variability of precipitation stable oxygen isotope values from two different sampling networks in Europe. To assess the performance of the models, mean squared error (MSE), nonparametric Kling Gupta efficiency (KGE), absolute differences and relative mean absolute error metrics were employed. It was found that the combination of the different regression techniques with Random Forest can produce estimations with comparable accuracy in terms of descending order of overall average MSE, MRRF: 2.61, RK: 2.77, RERF: 2.99, RF: 3.08. The best performing combined random forest model variant (MRRF) outperformed regression kriging in terms of a hybrid error metric (KGE) by 7.5%. Sequential random rarefying the station networks showed that machine-learning methods are more capable of maintaining high prediction accuracy even with fewer input data. This can be a great advantage when a suitable method is needed to predict the stable isotope composition of precipitation for large spatial domains where the spatial density of data stations shows large differences. [ABSTRACT FROM AUTHOR]

Extreme peak runoff forecasting is still a challenge in hydrology. In fact, the use of traditional physically-based models is limited by the lack of sufficient data and the complexity of the inner hydrological processes. Here, we employ a Machine Learning technique, the Random Forest (RF) together with a combination of Feature Engineering (FE) strategies for adding physical knowledge to RF models and improving their forecasting performances. The FE strategies include precipitation-event classification according to hydrometeorological criteria and separation of flows into baseflow and directflow. We used ∼ 3.5 years of hourly precipitation information retrieved from two near-real-time satellite precipitation databases (PERSIANN-CCS and IMERG-ER), and runoff data at the outlet of a 3391-km2 basin located in the tropical Andes of Ecuador. The developed models obtained Nash-Sutcliffe efficiencies varying from 0.86 to 0.59 for lead times between 1 and 6 h. The best performances were obtained for peak runoffs triggered by short-extension precipitation events (<50 km2) where infiltration- or saturation-excess runoff responses are well learned by the RF models. Conversely, the forecasting difficulty is associated with extensive precipitation events. For such conditions, a deeper characterization of the biophysical characteristics of the basin is encouraged for capturing the dynamic of directflow across multiple runoff responses. All in all, the potential to employ near-real-time satellite precipitation and the use of FE strategies for improving RF forecasting provides hydrologists with new tools for real-time runoff forecasting in remote or complex regions. • Near-real-time satellite-precipitation products were tested for peak runoff forecasting. • The addition of feature engineering strategies to machine learning models improved their forecasting efficiencies. • The methods proposed in this study can be used for runoff forecasting in remote or complex areas. [ABSTRACT FROM AUTHOR]

[Display omitted] • Fatigue life of bending polymer films under cyclic bending was measured with various test conditions. • The bending fatigue life of crystalline and amorphous polymer films was successfully predicted by a single machine learning model (random forest). • Important factors in the fatigue life of bending polymer films were estimated from the calculation of feature importance. The prediction of the fatigue life of polymer film substrates under cyclic bending plays an important role in designing durable flexible devices. Here, the fatigue life of bending polymer film was predicted using machine learning; the required data were collected via fatigue tests under different test conditions. Machine-learning models (linear regression and random forest regression) were constructed using these collected data. The random forest model predicted the fatigue life with a mean absolute percentage error of 22.3% within 1 min. Such accurate and efficient fatigue life predictions can contribute toward the development of flexible devices. [ABSTRACT FROM AUTHOR]

• Forecast electricity demand considering residents' psychological preferences. • Stationarity time series are established to avoid the impact of irregular factors. • Personalizes exogenous variables, combines correlations and importance. • A hybrid machine learning based model is built to improve forecasting performance. • Multiple implications are provided for the government and institutions. Faced with the explosive growth of population and rapid development of industrialization, household electricity demand (HED) has recently become an essential sector of energy demand. Very-short-term forecasting of HED is closely related to monitoring insecurity, detecting emergency and potential crisis. Consider challenges triggered by the characteristics of time series, firstly, a phase of data pre-processing with stationarity of time series is constructed to manage irregular trend factors. Subsequently, in the feature selection phase, a two-layer feature selection method is proposed with the combination of random forest and autocorrelation analysis to select the most relevant features, which simultaneously considers correlations and importance of input lagged variables as well as exogenous variables. After that, hybrid HED forecasting methods are constructed with incorporating multiple machine learning techniques in the forecasting phase, namely k-nearest neighbor, support vector regression , random forest and multilayer perceptron, using their respective advantages of managing different situations. As residents' psychological preferences and calendar variables highly affect future HED, this paper intends to add and test their effects on HED. Finally, the proposed methods are examined on a 10-min-ahead horizon application, whose forecasting results, comparison analyses and an extended experiment are given to illustrate their effectiveness, after which some implications are provided to show the insights gained from the results. [ABSTRACT FROM AUTHOR]

• A machine learning (ML) framework for fatigue life prediction is proposed. • Extraction of sensitive features weakens the effect of causality among features. • Physics-informed ML model for fatigue life prediction is constructed. • ML framework are employed to predict the fatigue life of AlSi10Mg alloy. Additive manufacturing (AM) process-induced defects make the fatigue life prediction of AM-built parts challenging. A machine learning (ML) framework based on sensitive features and continuous damage mechanics (CDM) herein is proposed to predict the fatigue life of AM-built parts. The sensitive features are extracted to blunt the disturbing effect of causality among the features. The CDM theory considering AM parameters is conducive to constructing a physics-informed ML model. This work employs support vector machines and random forests to predict the fatigue life of AM-built AlSi10Mg alloy. The results demonstrate that the physical knowledge-guided ML model using sensitive features exhibits better performance of fatigue life prediction. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning methods were employed to predict bed expansion ratio. • XGBoost model has good regression and prediction ability with R2 higher than 0.97. • The relative importance of operating variables on bed expansion ratio was obtained. • Particle size has the strongest effect on bed expansion ratio. Liquid–solid fluidizations are mainly affected by particle density, particle size, particle shape, superficial velocity, liquid viscosity, etc. Owing to these complex high-dimensional impact processes, rapid and accurate prediction of fluidization characteristics is necessary but, still challenging. In this study, machine learning models were developed and employed to predict bed expansion ratio. Machine learning models, namely, linear, random forest, and XGBoost regression models, were trained using a measured 183- bed expansion ratio dataset under different operating conditions. By comparison, the XGBoost model had strongest regression and prediction ability with R2 higher than 0.97 and the predictions were very time-saving with calculation time less than 1 s. Finally, the bed expansion characteristics were analyzed and the operating conditions were evaluated by the developed model. The relative importance of operating conditions on bed expansion ratios was particle size > particle density > superficial velocity. [ABSTRACT FROM AUTHOR]

• Forecasting the short-term daily streamflow is an essential and complex task in hydrology. • A novel stacked machine learning algorithm is compared to a powerful deep learning model. • The stacked model combines Random Forest and Multilayer Perceptron through Elastic Net. • The deep learning model is based on bidirectional Long Short-Term memory networks. • The two models showed comparable performance, strongly affected by the forecast horizon. Prediction of river flow rates is an essential task for both flood protection and optimal water resource management. The high uncertainty associated with basin characteristics, hydrological processes, and climatic factors affecting river flows make streamflow prediction a very challenging problem. These reasons, together with the increasingly wide availability of data relating to flow rates and rainfall, frequently lead to a preference for data-driven models over physically based or conceptual forecasting models. This study shows the results of an in-depth comparison between two different daily streamflow prediction models: a novel simpler model based on the stacking of the Random Forest and Multilayer Perceptron algorithms, using the Elastic Net algorithm as meta -learner, and a more complex model based on bi-directional Long Short-Term Memory (LSTM) networks. Bayesian optimization was employed in selecting hyperparameters. The two prediction models were compared through the analysis of four different case studies: the Bacchiglione River, the Raccoon River, the Wilson River, and the Trent River. The two models showed comparable forecasting capabilities. In the most favourable cases, both models demonstrated very high accuracy (R2 greater than 0.93, Mean absolute percentage error approximately equal to 10%). The stacked model outperformed the bi-directional LSTM network model in several cases in predicting peak flow rates but was less accurate in forecasting low flow rates. In addition, its computation times are significantly shorter. The prediction accuracy of both models decreased as the forecast horizon increased. The length of the time series plays an essential role in developing models with satisfactory forecasting capabilities. In this study, the effectiveness of forecasting models was not influenced by the river regime. However, a high variance of the input dataset and a large number of outliers in the time series can reduce the accuracy of prediction models. [ABSTRACT FROM AUTHOR]

Main aspect of this study is to explore the possible use of Machine learning based approaches in the petroleum industry. The fourth industrial revolution is being ushered in by advances in computer power. The oil and gas sector contains a huge amount of produced oil data from wells that can be utilised to create data-driven models which can validate empirical correlations. This validation is vital in the current context because the oil market fluctuates a lot, thus making efficient judgments based on insights supplied by data science and machine learning algorithms are valuable. Artificial Neural Network (ANN), Random Forest Regressor (RF) and Gradient Boosting Regressor (GB) were used in this study to forecast daily oil production using available production parameters. In this study, the production data from Equinor's Volve Field was utilised. The obtained data was filtered, and then it was applied to mentioned algorithms with fine tuned hyperparameters. The data was split into training data (70%) and test data (30%). The model was constructed on training data then it was validated through test data. Seven parameters were chosen on the basis of magnitude of correlation coefficient. The result showed that even with the cases of several shut-ins, the hydrocarbon forecasting method for well 159-F-1C predicted production rates with a coefficient of determination (R2score) value greater or equal to 0.90. For performance analysis, coefficient of determination (R2score), mean squared error (MSE) and mean absolute error (MAE) were used for the selected wells. The outcome showed that prediction of hydrocarbon production for well 15/9-F-1 C ANN model showed the best performance (R2 score = 0.9, MAE = 20.52 and MSE = 1268) and for well 15/9-F-12 RF model shows the best performance (R2 score = 0.98, MAE = 84.95 and MSE = 37356.07). This novel approach could be trained on any dataset to assist in oil forecasting. • If sufficient historical data of any field is available it is possible to construct a model that can predict hydrocarbon production. • Data-driven approaches can predict hydrocarbon production for unconventional resources. • Machine learning algorithms cannot replace the conventional methods, but can assist conventional methods in improving efficiency and accuracy. • Different machine learning algorithms should be trained and evaluated on the dataset for selecting the best algorithm. • There is no one specific model for production prediction as no machine learning algorithm is superior. [ABSTRACT FROM AUTHOR]

• In this paper, a framework consisting of a pipeline of methodologies is proposed for short-term prediction of broken rails. The effectiveness of the proposed framework is validated using data from one Class I railroad. • Feature-based track segmentation is proposed to delineate the railroad network, which is able to improve the performance of the proposed model by reducing variations in data records. • This paper extracts more features and uses them as input variables for the machine learning models. While previous studies did not include them, the feature importance identifies the top influencing variables, such as minimum temperature and crossing angles of turnouts, for broken rail prediction. Predicting the occurrence of broken rails has safety and economic benefits, and reduces accidents and service disruptions. This paper aims to build a data-driven model for broken rail prediction using data related to infrastructure, operations, inspections, and weather conditions from 2013 to 2019. The railroad data was provided by one major Class I U.S. railroad. The weather condition data was collected from the National Oceanic and Atmospheric Administration (NOAA). Based on time-series data partitioning, three different machine learning models are developed for predicting broken rail occurrence one month in advance. The selected models, including logistic regression, random forests, and gradient boosting, are trained using data from 2013 to 2018. The performance of three trained models are evaluated using the data from 2019. The relationship between the percentage of the network scanned and the percentage of broken rails found is used to identify locations that are more prone to broken rails. The findings of this study show that the gradient boosting model performs better than the other two methods for our datasets. The model also identifies that the number of detected rail defects within last 365 days, minimum ambient temperature of the last 30 days, days from the last broken rail, segment length, traffic density and other factors have significant influences on prediction results. Using this model, 40.4% of broken rails can be successfully predicted one month in advance when focusing on 10% of the railroad network scanned. The model can potentially be used to prioritize inspection and maintenance activities on broken-rail-prone locations, and thus to further improve infrastructure safety given budgetary constraints. [ABSTRACT FROM AUTHOR]

[Display omitted] • Predictive model for average dissolved oxygen (DO) levels is robust and reliable. • Maintaining optimal flocculation level in pond helps in healthy growth of shrimp. • Feature selection via Mutual Information is highly accurate for DO prediction. • Including Meteorological parameters in model development improves model accuracy. The present study approaches towards the feasibility of prediction and forecasting of dissolved oxygen (DO) and biofloc amount using the state of art machine learning algorithms in a shrimp culture system. The study was carried out considering; average DO and biofloc amount as target parameters in a shrimp culture system. There were seventeen numbers of culture and meteorological parameters considered and three different feature selection techniques used to create twelve different data subsets for model development. The model development was carried out using three popular machine learning algorithms viz., Random Forest, Adaboost and Deep neural network. The totals of thirty-six different models were obtained and their accuracies were evaluated with seven model validation tests and results were obtained and discussed. Out of thirty-six models, Random Forest technique applied model for prediction of dissolved oxygen with combined culture and meteorological parameters (R2value – 0.709, prediction accuracy – 98.26%, score – 0.7381) was found to be the best one for predictive model development. Moreover, exploratory data analysis was carried out for prediction and a framework for prediction of DO and biofloc amount in a shrimp based biofloc culture system was developed. The dissolved oxygen was found to be more robust in the predictive model development. The Intelligent Framework was developed based on the study conducted to understand and carryout prediction in farming system in a scientific manner. The developed framework can help the literate farmers and new entrants of shrimp farming to devise their own prediction models suitable for the farming conditions. [ABSTRACT FROM AUTHOR]

This paper analyzes the performance of three types of statistical models and a well-known physics-based model for forecasting the wave energy flux. The forecasts are run over horizons of 1–24 h at five buoys located in the Bay of Biscay. The data resolution is hourly. The 1999–2005 timeframe is used to train the models. The forecasts are run and evaluated over the six-year period from 2006 to 2012. The statistical forecasting models use three techniques: analogues, random forests (a machine learning algorithm) and a combination of the two. The physics model is the Wave Model (WAM). The forecasts are compared at a 95% confidence level with the simplest prediction – Persistence – and also with the nearest grid point of the WAM forecasts. Over horizons between 3 and 16–19 h at locations near the coast (where wave farms may be installed), the random forests models outperform the others, including WAM and Persistence. These models exploit the inherent predictability associated with the strong autocorrelation present in ocean energy values. The additional prognostic capabilities that random forests models provide over Persistence, are due to their ability to sucessfully incorporate the information that both, atmospheric and sea-state variables provide. [ABSTRACT FROM AUTHOR]

• We study methods for quantifying predictive uncertainty in machine learning. • Conformal prediction is a technique for forming distribution-free prediction intervals. • Nested conformal is an intuitive rephrasing of conformal in terms of nested sets. • Our algorithm QOOB combines quantile random forests with the leave-one-out jackknife. • QOOB achieves state-of-the-art results for conformal prediction in regression. Conformal prediction is a popular tool for providing valid prediction sets for classification and regression problems, without relying on any distributional assumptions on the data. While the traditional description of conformal prediction starts with a nonconformity score, we provide an alternate (but equivalent) view that starts with a sequence of nested sets and calibrates them to find a valid prediction set. The nested framework subsumes all nonconformity scores, including recent proposals based on quantile regression and density estimation. While these ideas were originally derived based on sample splitting, our framework seamlessly extends them to other aggregation schemes like cross-conformal, jackknife+ and out-of-bag methods. We use the framework to derive a new algorithm (QOOB, pronounced cube) that combines four ideas: quantile regression, cross-conformalization, ensemble methods and out-of-bag predictions. We develop a computationally efficient implementation of cross-conformal, that is also used by QOOB. In a detailed numerical investigation, QOOB performs either the best or close to the best on all simulated and real datasets. [ABSTRACT FROM AUTHOR]