Stacking Algorithms Research Articles

Multi-Task Forecasting of the Realized Volatilities of Agricultural Commodity Prices

Motivated by the comovement of realized volatilities (RVs) of agricultural commodity prices, we study whether multi-task forecasting algorithms improve the accuracy of out-of-sample forecasts of 15 agricultural commodities during the sample period from July 2015 to April 2023. We consider alternative multi-task stacking algorithms and variants of the multivariate Lasso estimator. We find evidence of in-sample predictability but scarce evidence that multi-task forecasting improves out-of-sample forecasts relative to a classic univariate heterogeneous autoregressive (HAR)-RV model. This lack of systematic evidence of out-of-sample forecasting gains is corroborated by extensive robustness checks, including an in-depth study of the quantiles of the distributions of the RVs and subsample periods that account for increases in the total spillovers among the RVs. We also study an extended model that features the RVs of energy commodities and precious metals, but our conclusions remain unaffected. Besides offering important lessons for future research, our results are interesting for financial market participants, who rely on accurate forecasts of RVs when solving portfolio optimization and derivatives pricing problems, and policymakers, who need accurate forecasts of RVs when designing policies to mitigate the potential adverse effects of a rise in the RVs of agricultural commodity prices and the concomitant economic and political uncertainty.

Multiparametric MRI-based radiomics combined with 3D deep transfer learning to predict cervical stromal invasion in patients with endometrial carcinoma.

To develop and compare various preoperative cervical stromal invasion (CSI) prediction models, including radiomics, three-dimensional (3D) deep transfer learning (DTL), and integrated models, using single-sequence and multiparametric MRI. Data from 466 early-stage endometrial carcinoma (EC) patients from three centers were collected. Radiomics models were constructed based on T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) mapping, contrast-enhanced T1-weighted imaging (CE-T1WI), and four combined sequences as well as 3D DTL models. Two integrated models were created using ensemble and stacking algorithms based on optimal radiomics and DTL models. Model performance and clinical benefits were assessed using area under the curve (AUC), decision curve analysis (DCA), net reclassification index (NRI), integrated discrimination index (IDI), and the Delong test for model comparisons. Multiparametric MRI models were superior to single-sequence models for radiomics or DTL models. Ensemble and stacking integrated models displayed excellent performance. The stacking model had the highest average AUC (0.908) and accuracy (0.883) in external validation groups 1 and 2 (AUC = 0.965 and 0.851, respectively) and emerged as the best predictive model for CSI. All models significantly outperformed the radiologist (P < 0.05). In terms of net benefits, all models demonstrated favorable outcomes in DCA, NRI, and IDI, with the stacking model yielding the highest net benefit. Multiparametric MRI-based radiomics combined with 3D DTL can be used to noninvasively predict CSI in EC patients with greater diagnostic accuracy than the radiologist. Stacking integrated models showed significant potential utility in predicting CSI. Which helps to provide new treatment strategy for clinicians to treat early-stage EC patients.

The drug loading capacity prediction and cytotoxicity analysis of metal–organic frameworks using stacking algorithms of machine learning

Metal–organic frameworks (MOFs) have shown excellent performance in the field of drug delivery. Despite the synthesis of a vast array of MOFs exceeding 100,000 varieties, certain formulations have exhibited suboptimal performance characteristics. Therefore, there is a pressing need to enhance their efficacy by identifying MOFs with superior drug loading capacities and minimal cytotoxicity, which can be achieved through machine learning (ML). In this study, a stacking regression model was developed to predict drug loading capacity and cytotoxicity of MOFs using datasets compiled from various literature sources. The model exhibited exceptional predictive capabilities, achieving R2 values of 0.907 for drug loading capacity and 0.856 for cytotoxicity. Furthermore, various model interpretation methods including partial dependence plots, individual conditional expectation, Shapley additive explanation, decision tree, random forest, CatBoost Regressor, and light gradient-boosting machine were employed for feature importance analysis. The results revealed that specific metal atoms such as Zn, Cr, Fe, Zr, and Cu significantly influenced the drug loading capacity and cytotoxicity of MOFs. Through model validation encompassing experimental validation and computational verification, the reliability of the model was thoroughly established. In general, it is a good practice to use ML methods for predicting drug loading capacity and cytotoxicity analysis of MOFs, guiding the development of future property prediction methods for MOFs.

A comparative study of explainable ensemble learning and logistic regression for predicting in-hospital mortality in the emergency department

This study addresses the challenges associated with emergency department (ED) overcrowding and emphasizes the need for efficient risk stratification tools to identify high-risk patients for early intervention. While several scoring systems, often based on logistic regression (LR) models, have been proposed to indicate patient illness severity, this study aims to compare the predictive performance of ensemble learning (EL) models with LR for in-hospital mortality in the ED. A cross-sectional single-center study was conducted at the ED of Imam Reza Hospital in northeast Iran from March 2016 to March 2017. The study included adult patients with one to three levels of emergency severity index. EL models using Bagging, AdaBoost, random forests (RF), Stacking and extreme gradient boosting (XGB) algorithms, along with an LR model, were constructed. The training and validation visits from the ED were randomly divided into 80% and 20%, respectively. After training the proposed models using tenfold cross-validation, their predictive performance was evaluated. Model performance was compared using the Brier score (BS), The area under the receiver operating characteristics curve (AUROC), The area and precision–recall curve (AUCPR), Hosmer–Lemeshow (H–L) goodness-of-fit test, precision, sensitivity, accuracy, F1-score, and Matthews correlation coefficient (MCC). The study included 2025 unique patients admitted to the hospital’s ED, with a total percentage of hospital deaths at approximately 19%. In the training group and the validation group, 274 of 1476 (18.6%) and 152 of 728 (20.8%) patients died during hospitalization, respectively. According to the evaluation of the presented framework, EL models, particularly Bagging, predicted in-hospital mortality with the highest AUROC (0.839, CI (0.802–0.875)) and AUCPR = 0.64 comparable in terms of discrimination power with LR (AUROC (0.826, CI (0.787–0.864)) and AUCPR = 0.61). XGB achieved the highest precision (0.83), sensitivity (0.831), accuracy (0.842), F1-score (0.833), and the highest MCC (0.48). Additionally, the most accurate models in the unbalanced dataset belonged to RF with the lowest BS (0.128). Although all studied models overestimate mortality risk and have insufficient calibration (P > 0.05), stacking demonstrated relatively good agreement between predicted and actual mortality. EL models are not superior to LR in predicting in-hospital mortality in the ED. Both EL and LR models can be considered as screening tools to identify patients at risk of mortality.

Development of ensemble learning models for prognosis of hepatocellular carcinoma patients underwent postoperative adjuvant transarterial chemoembolization.

Postoperative adjuvant transarterial chemoembolization (PA-TACE) has been increasing widely used to improve the prognosis of hepatocellular carcinoma (HCC) patients. However, clinical outcomes vary from patient to patient, which calls for individualized prognostic prediction and early management. A total of 274 HCC patients who underwent PA-TACE were enrolled in this study. The prediction performance of five machine learning models was compared and the prognostic variables of postoperative outcomes were identified. Compared with other machine learning models, the risk prediction model based on ensemble learning strategies, including Boosting, Bagging, and Stacking algorithms, presented better prediction performance for overall mortality and HCC recurrence. Moreover, the results showed that the Stacking algorithm had relatively low time consumption, good discriminative ability, and the best prediction performance. In addition, according to time-dependent ROC analysis, the ensemble learning strategies were found to perform well in predicting both OS and RFS for the patients. Our study also found that BCLC Stage, hsCRP/ALB and frequency of PA-TACE were relatively important variables in both overall mortality and recurrence, while MVI contributed more to the recurrence of the patients. Among the five machine learning models, the ensemble learning strategies, especially the Stacking algorithm, could better predict the prognosis of HCC patients following PA-TACE. Machine learning models could also help clinicians identify the important prognostic factors that are clinically useful in individualized patient monitoring and management.

Rockburst Intensity Classification Prediction Based on Multi-Model Ensemble Learning Algorithms

Rockburst is a common and huge hazard in underground engineering, and the scientific prediction of rockburst disasters can reduce the risks caused by rockburst. At present, developing an accurate and reliable rockburst risk prediction model remains a great challenge due to the difficulty of integrating fusion algorithms to complement each other’s strengths. In order to effectively predict rockburst risk, firstly, 410 sets of valid rockburst data were collected as the original data set in this paper, which was used to process these rockburst cases by the SMOTE oversampling method. Then, four integrated algorithms and eight basic algorithms were selected, which were optimized by hyperparameters and five-fold cross-validation and combined with the random search grid method, thus improving the classification performance of these algorithms. Third, the stacking integration algorithm, which was combined with the principles of various machine learning algorithms and the characteristics of the rockburst cases, integrated the optimization of rockburst algorithms with reference to four combinatorial strategies. Further, we adopted the voting integration algorithm, chose multiple combination schemes, and referred to the weighted fusion of accuracy, F1 score, recall, precision, and cv-mean as the weight values, and the optimal model for rockburst risk prediction was obtained. Finally, using the 35 generated stacking integration algorithms and 18 voting integration algorithms, the optimal model in the fusion strategy was selected and the traditional integration algorithm model was analyzed on the basis of different sample combinations of the models. The results showed that the prediction performance of stacking and voting integration algorithms was mostly better than the ordinary machine-learning performance, and the selection of appropriate fusion strategies could effectively improve the performance of rockburst prediction for ensemble learning algorithms.

An interpretable machine learning approach for engineering change management decision support in automotive industry

As an essential part of the Product Life Cycle (PLC), the time-to-market of products is influenced by Engineering Change Management (ECM) processes. An Engineering Change (EC) is part of a formal process in industry to describe, rationalize, determine, release components for final production or make changes to already released design. It includes information about shape, functionality, production location, cost, and other relevant data entries. The duration from creation to the approval of a change request can take weeks or even months, without apparent reasons for the bottleneck. In addition, changes to one component can lead to unexpected chain reactions to other components. Therefore, identifying impacts of changes is challenging for all Original Equipment Manufacturers (OEMs). To address the above challenges, the authors have developed and built a machine learning-based decision support solution in this article. Community detection and stacking algorithms were applied to build more robust models. Impacts and lead time of Engineering Change Requests (ECRs) are predicted and explained by Local Interpretable Model-agnostic Explanations (LIME). A case study was conducted on real-world data from an automotive company. After evaluation with industry experts, the solution approach was proved to have positive contributions to increasing the quality, efficiency, and transparency of the existing ECM processes.

The prediction of product return rates with ensemble machine learning algorithms

There may not always be actual data available for planning. Predicted data are used especially for future planning. Due to errors in such planning based on prediction, many products enter the reverse logistics network without completing the shelf life. Especially in textile sector, because of fashion, it is the most important point of planning to be able to make accurate estimates in order to avoid unnecessary resource utilization and to provide minimum cost. It is difficult to establish a mathematical model because the prediction problems in real life have multivariate structure and unknown parameters. Most of the studies in literature have been based on time series prediction. But due to changing fashion and demands of consumers, there are significant differences between demand forecasts and real data. So, in the problems with unknown parameters and multivariate structure, Ensemble Machine Learning (EML) methods are preferred recently because they give more accurate results than other prediction methods. Unlike other studies, the product return rate in textile sector has been predicted with the Stacking and Vote algorithms from EML methods in this paper. In this direction, it is aimed to concentrate on the returns of the products sold with the preferences of the customers and to predict the returns more accurately. In this way, the consumer information obtained as a result of the analyzes can provide more accurate planning in avoiding unnecessary production, transportation and storage activities, reducing costs, resource utilization and environmental pollution. In addition, it is one of the main aims of the study to contribute to the literature by determining the parameters that can be used in predicting the return rates. Highest performance results were obtained with Stacking algorithm. The obtained results were given comparatively and the correlation coefficient of 86.07% was reached.

Estimation of solubility of acid gases in ionic liquids using different machine learning methods

Ionic liquids (ILs) can capture acid gases that damaged the environment. Due to the properties of low-cost and non-toxic, machine learning can be used to screen ILs for gas absorption. To find the most suitable machine learning method for estimating gas absorption in ILs, 12 different machine learning algorithms are used to train models to estimate CO2 and H2S solubility in different ILs. Temperature (T), pressure (P), molecular weight (Mw), critical temperature (Tc), and critical pressure (Pc) of solutions are used as the input variables; Solubility is used as the output variable in the model training. Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Correlation Coefficient (R2) are used to evaluate the models. Stacking algorithm has the most accurate model in IL- CO2 system, with MSE, RMSE, MAE, and R2 of 0.001, 0.025, 0.018, and 0.969 respectively on average. Voting algorithm performs best in IL-H2S system; the four averaged metrics are 0.002, 0.032, 0.024, and 0.934 accordingly. By combining different algorithms, Voting and Stacking algorithms can balance out each model's weakness and produce a more accurate model. Stacking and Voting algorithms can be considered as a promising candidate for the estimation of acid gases solubility in ionic liquids.

SPI: Automated Identification of Security Patches via Commits

Security patches in open source software, providing security fixes to identified vulnerabilities, are crucial in protecting against cyber attacks. Security advisories and announcements are often publicly released to inform the users about potential security vulnerability. Despite the National Vulnerability Database (NVD) publishes identified vulnerabilities, a vast majority of vulnerabilities and their corresponding security patches remain beyond public exposure, e.g., in the open source libraries that are heavily relied on by developers. As many of these patches exist in open sourced projects, the problem of curating and gathering security patches can be difficult due to their hidden nature. An extensive and complete security patches dataset could help end-users such as security companies, e.g., building a security knowledge base, or researcher, e.g., aiding in vulnerability research. To efficiently curate security patches including undisclosed patches at large scale and low cost, we propose a deep neural-network-based approach built upon commits of open source repositories. First, we design and build security patch datasets that include 38,291 security-related commits and 1,045 Common Vulnerabilities and Exposures (CVE) patches from four large-scale C programming language libraries. We manually verify each commit, among the 38,291 security-related commits, to determine if they are security related. We devise and implement a deep learning-based security patch identification system that consists of two composite neural networks: one commit-message neural network that utilizes pretrained word representations learned from our commits dataset and one code-revision neural network that takes code before revision and after revision and learns the distinction on the statement level. Our system leverages the power of the two networks for Security Patch Identification. Evaluation results show that our system significantly outperforms SVM and K-fold stacking algorithms. The result on the combined dataset achieves as high as 87.93% F1-score and precision of 86.24%. We deployed our pipeline and learned model in an industrial production environment to evaluate the generalization ability of our approach. The industrial dataset consists of 298,917 commits from 410 new libraries that range from a wide functionalities. Our experiment results and observation on the industrial dataset proved that our approach can identify security patches effectively among open sourced projects.

Short-Term Wind Power Prediction Approach Based On Bayesian Optimization and Ensemble Learning

In wind energy studies, predicting the short-term energy generation amount for wind power plants and determining the production offer to be placed on the market play an important role. In this study an hourly short-term wind power estimation of a wind turbine located in Turkey with an installed power of 3600 kW has been made. Estimation results were evaluated on a seasonal and annual basis. New hybrid models have been developed for short-term wind power prediction, consisting of Bayesian Optimization (BO), Support Vector Regression (SVR), Gaussian Process Regression (GPR), Decision Tree (DT), stacking, and bagging algorithms. In the proposed prediction approach, it is aimed to reduce prediction errors by combining different regression algorithms with the BO method and ensemble algorithms. Unlike other wind prediction studies, BO was used for the first time in the hyperparameter selection of the regression algorithms selected as the basic learner in the study. Bayesian optimized decision tree (BO-DT) with the lowest error values among the base learners, and Bayesian optimized gaussian process regression (BO-GPR) combined with bagging and stacking. The efficiency of ensemble learning algorithms was measured by the statistical measurement methods Normalized Absolute Mean Error (NMAE), Normalized Root of Mean Squares Error (NRMSE), and determination coefficient (R 2 ). According to the results, the bagging method created with the BO-DT took the annual average NRMSE, NMAE, R 2 criteria of 11.045%, 4.880%, 0.899, respectively, and the model with the best performance was selected in terms of both annual and seasonal results.

A stack-based set inversion model for smart water, carbon and ecological assessment in urban agglomerations

Footprint evaluation is an important tool for assessing the appropriation of ecological assets, GHG emissions, freshwater consumption parameters, etc., within a specified region. However, traditional evaluation of footprints for mega cities or urban agglomerations requires overmuch different types of high-quality data. There is a great need of seeking a smart model/approach with declined data requirements for evaluation of footprints where part of data can hardly be accessed. Here we propose a new ensemble inversion model (EIM) based on integrated multitask machine learning (MML) and multi-modeling stacking (MMS) algorithms for smart evaluation and prediction of water, carbon and ecological footprints. The accuracy and generalization capability of the model are illustrated through three largest urban agglomerations in the middle reaches of the Yangtze River (MRYR). The testing results show that the EIM achieves similar prediction performance compared to traditional footprints calculation methods (R2 = 0.91, RMSE = 0.18, MAE = 0.11), yet greatly reduces the amount of required data by approximately 80%. Moreover, the accuracy of the EIM is improved by more than 20%, compared with other models using a single inversion algorithm. The modeling results also show that 1) water, carbon and ecological footprints are significantly positively correlated, and 2) an annual increase of 4.8% can be found in terms of the urban environmental pressure index (UEPI), and its projection is even less optimistic for the future.

Spatial Modeling of Asthma-Prone Areas Using Remote Sensing and Ensemble Machine Learning Algorithms

In this study, asthma-prone area modeling of Tehran, Iran was provided by employing three ensemble machine learning algorithms (Bootstrap aggregating (Bagging), Adaptive Boosting (AdaBoost), and Stacking). First, a spatial database was created with 872 locations of asthma patients and affecting factors (particulate matter (PM10 and PM2.5), ozone (O3), sulfur dioxide (SO2), carbon monoxide (CO), nitrogen dioxide (NO2), rainfall, wind speed, humidity, temperature, distance to street, traffic volume, and a normalized difference vegetation index (NDVI)). We created four factors using remote sensing (RS) imagery, including air pollution (O3, SO2, CO, and NO2), altitude, and NDVI. All criteria were prepared using a geographic information system (GIS). For modeling and validation, 70% and 30% of the data were used, respectively. The weight of evidence (WOE) model was used to assess the spatial relationship between the dependent and independent data. Finally, three ensemble algorithms were used to perform asthma-prone areas mapping. According to the Gini index, the most influential factors on asthma occurrence were distance to the street, NDVI, and traffic volume. The area under the curve (AUC) of receiver operating characteristic (ROC) values for the AdaBoost, Bagging, and Stacking algorithms was 0.849, 0.82, and 0.785, respectively. According to the findings, the AdaBoost algorithm outperforms the Bagging and Stacking algorithms in spatial modeling of asthma-prone areas.

Delineating Smallholder Maize Farms from Sentinel-1 Coupled with Sentinel-2 Data Using Machine Learning

Rural communities rely on smallholder maize farms for subsistence agriculture, the main driver of local economic activity and food security. However, their planted area estimates are unknown in most developing countries. This study explores the use of Sentinel-1 and Sentinel-2 data to map smallholder maize farms. The random forest (RF), support vector (SVM) machine learning algorithms and model stacking (ST) were applied. Results show that the classification of combined Sentinel-1 and Sentinel-2 data improved the RF, SVM and ST algorithms by 24.2%, 8.7%, and 9.1%, respectively, compared to the classification of Sentinel-1 data individually. Similarities in the estimated areas (7001.35 ± 1.2 ha for RF, 7926.03 ± 0.7 ha for SVM and 7099.59 ± 0.8 ha for ST) show that machine learning can estimate smallholder maize areas with high accuracies. The study concludes that the single-date Sentinel-1 data were insufficient to map smallholder maize farms. However, single-date Sentinel-1 combined with Sentinel-2 data were sufficient in mapping smallholder farms. These results can be used to support the generation and validation of national crop statistics, thus contributing to food security.

Cloud based ensemble machine learning approach for smart detection of epileptic seizures using higher order spectral analysis.

The present paper proposes a smart framework for detection of epileptic seizures using the concepts of IoT technologies, cloud computing and machine learning. This framework processes the acquired scalp EEG signals by Fast Walsh Hadamard transform. Then, the transformed frequency-domain signals are examined using higher-order spectral analysis to extract amplitude and entropy-based statistical features. The extracted features have been selected by means of correlation-based feature selection algorithm to achieve more real-time classification with reduced complexity and delay. Finally, the samples containing selected features have been fed to ensemble machine learning techniques for classification into several classes of EEG states, viz. normal, interictal and ictal. The employed techniques include Dagging, Bagging, Stacking, MultiBoost AB and AdaBoost M1 algorithms in integration with C4.5 decision tree algorithm as the base classifier. The results of the ensemble techniques are also compared with standalone C4.5 decision tree and SVM algorithms. The performance analysis through simulation results reveals that the ensemble of AdaBoost M1 and C4.5 decision tree algorithms with higher-order spectral features is an adequate technique for automated detection of epileptic seizures in real-time. This technique achieves 100% classification accuracy, sensitivity and specificity values with optimally small classification time.

A migration-based location method using improved waveform stacking for microseismic events in a borehole system

The migration-based microseismic event location methods using waveform stacking algorithms are widely used for hydro-fracturing monitoring. These methods have the advantage of not requiring the accurate first arrival time around a detected event, which is more suitable for noisy data than classical travel time-based methods. However, accuracy of these methods can be affected under the condition of relatively low signal-to-noise ratio (SNR). Therefore, in order to enhance the location accuracy of microseismic events in a borehole system, we have proposed a migration-based location method using improved waveform stacking with polarity correction based on a master-event technique, which optimizes the combination way of P- and S-wave waveform stacking. This method can enhance the convergence of the objective function and the location accuracy for microseismic events as compared to the conventional waveform stacking. The proposed method has been successfully tested by using synthetic data example and field data recorded from one downhole monitoring well. Our study clearly indicates that the presented method is more viable and stable under low SNR.

Bus Travel Time Prediction Based on Ensemble Learning Methods

Buses have become indispensable in urban transportation systems, especially in developing countries. The prediction of bus travel time can provide essential information for passengers to coordinate their trip plans. By combining several prediction algorithms, ensemble learning methods have shown great potential for improving prediction accuracy in many research fields. In this article, ensemble learning methods are used to predict bus travel times. First, a novel feature-selection algorithm, Boruta, is introduced to select the appropriate input features for predicting bus travel time. The algorithm can quantify the importance of each feature and identify those that are important for the prediction. Second, we illustrate the ensemble learning methods in detail, including the bagging, boosting, and stacking methods. The representative algorithm of each category of methods is presented and utilized to study the prediction problem. Finally, a case study is conducted based on real-world data. Twenty original features are analyzed using the Boruta algorithm, and two are filtered out. Besides the ensemble learning algorithms, we also choose some other classical algorithms to predict the bus travel time. The results show that the boosting and stacking algorithms outperform other algorithms in terms of prediction accuracies.

A cross-entropy based stacking method in ensemble learning

Stacking is one of the major types of ensemble learning techniques in which a set of base classifiers contributes their outputs to the meta-level classifier, and the meta-level classifier combines them so as to produce more accurate classifications. In this paper, we propose a new stacking algorithm that defines the cross-entropy as the loss function for the classification problem. The training process is conducted by using a neural network with the stochastic gradient descent technique. One major characteristic of our method is its treatment of each meta instance as a whole with one optimization model, which is different from some other stacking methods such as stacking with multi-response linear regression and stacking with multi-response model trees. In these methods each meta instance is divided into a set of sub-instances. Multiple models apply to those sub-instances and each for a class label. There is no connection between different models. It is very likely that our treatment is a better choice for finding suitable weights. Experiments with 22 data sets from the UCI machine learning repository show that the proposed stacking approach performs well. It outperforms all three base classifiers, several state-of-the-art stacking algorithms, and some other representative ensemble learning methods on average.

How to avoid false interpretations of Sentinel-1A TOPSAR interferometric data in landslide mapping? A case study: recent landslides in Transdanubia, Hungary

It is a crucial issue to better understand the usability of Sentinel-1 satellites in geomorphologic applications, since Sentinel-1 and the Copernicus Program are considered to be the workhorse of Earth observation by the European Space Agency during the next decades. Yet, a very limited experience is available on the applicability of Sentinel-1 images in the detection and identification of surface deformations and especially landslide mapping and monitoring in densely vegetated (low-coherence) areas. Few Synthetic Aperture Radar images (not more than 20) are sufficient for a successful run of interferometric stacking algorithms. This number is really low compared to the tremendous data flow of Sentinel-1 images that are available for interferometric analysis nowadays. Despite the availability of acquisitions, only a few papers exist on the accuracy of Sentinel-1 data, signal-to-noise ratio and the value of the acquired imagery for geomorphologic interpretation. Two test sites and a control site—affected by active surface deformations—have been investigated using 40 Sentinel-1A images and conventional persistent scatterers (PSI) method. PSI results have been combined with the geomorphologic information of the studied sites. We verified that the given number of Sentinel-1A acquisitions provide a unique base for surface deformation recognition and mapping in low-coherence areas. We found that scatterers were corrupted by a strong noise if their line of sight (LOS) velocity was below ± 6–7 mm/year all over the three test sites, although noise can easily be reduced. Noise reduction was achieved by a significant increase of the length of time series, i.e., time range between the first and last image to reduce the effect of atmospheric phase screen (APS).

State of health prediction for lithium-ion batteries using multiple-view feature fusion and support vector regression ensemble

Lithium-ion batteries have been widely used in many electronic systems. Accurately estimating the state of health (SOH) of a lithium-ion battery is important for ensuring its safety and reliability. Among the various kinds of methods for predicting the SOH of lithium-ion batteries, machine-learning-based methods are the most popular. However, two common critical problems in machine-learning-based methods are extracting discriminative features and effectively utilizing the extracted features. In this study, we focused on solving these two issues. First, a sliding-window-based feature extraction technology (SWBFE) was designed to effectively extract features from different views in the discharge process of lithium-ion batteries. Second, we developed a multiple-view feature fusion with a support vector regression (SVR) ensemble strategy (MVFF-ESVR) for enhancing the performance in fusing multiple extracted features. The basic idea of MVFF-ESVR is to transform the feature-level fusion problem into a decision-level fusion problem. More specifically, for each feature, an SVR was modeled on the corresponding training set, and the AdaBoost and Stacking algorithms were utilized to incorporate multiple trained SVRs for generating two ensemble SVR models. By combining SWBFE with MVFF-ESVR, we further implemented two predictors, namely, Ada-TargetSOH and Sta-TargetSOH, for robust prediction of lithium-ion battery SOH. To evaluate the efficacy of the proposed predictors, we applied Ada-TargetSOH and Sta-TargetSOH on three types of lithium-ion battery datasets. The experimental results have demonstrated that our predictors outperform other existing lithium-ion battery SOH predictors.