support vector machine research paper

• Research article
• Open access
• Published: 22 March 2010

Application of support vector machine modeling for prediction of common diseases: the case of diabetes and pre-diabetes

• Tiebin Liu 1 ,
• Rodolfo Valdez 1 ,
• Marta Gwinn 1 &
• Muin J Khoury 1  

BMC Medical Informatics and Decision Making volume  10 , Article number:  16 ( 2010 ) Cite this article

51k Accesses

287 Citations

8 Altmetric

Metrics details

We present a potentially useful alternative approach based on support vector machine (SVM) techniques to classify persons with and without common diseases. We illustrate the method to detect persons with diabetes and pre-diabetes in a cross-sectional representative sample of the U.S. population.

We used data from the 1999-2004 National Health and Nutrition Examination Survey (NHANES) to develop and validate SVM models for two classification schemes: Classification Scheme I (diagnosed or undiagnosed diabetes vs. pre-diabetes or no diabetes) and Classification Scheme II (undiagnosed diabetes or pre-diabetes vs. no diabetes). The SVM models were used to select sets of variables that would yield the best classification of individuals into these diabetes categories.

For Classification Scheme I, the set of diabetes-related variables with the best classification performance included family history, age, race and ethnicity, weight, height, waist circumference, body mass index (BMI), and hypertension. For Classification Scheme II, two additional variables--sex and physical activity--were included. The discriminative abilities of the SVM models for Classification Schemes I and II, according to the area under the receiver operating characteristic (ROC) curve, were 83.5% and 73.2%, respectively. The web-based tool-Diabetes Classifier was developed to demonstrate a user-friendly application that allows for individual or group assessment with a configurable, user-defined threshold.

Conclusions

Support vector machine modeling is a promising classification approach for detecting persons with common diseases such as diabetes and pre-diabetes in the population. This approach should be further explored in other complex diseases using common variables.

Peer Review reports

A supervised machine learning method, the support vector machine (SVM) algorithm [ 1 ], has demonstrated high performance in solving classification problems in many biomedical fields, especially in bioinformatics [ 2 , 3 ]. In contrast to logistic regression, which depends on a pre-determined model to predict the occurrence or not of a binary event by fitting data to a logistic curve, SVM discriminates between two classes by generating a hyperplane that optimally separates classes after the input data have been transformed mathematically into a high-dimensional space. Because the SVM approach is data-driven and model-free, it may have important discriminative power for classification, especially in cases where sample sizes are small and a large number of variables are involved (high-dimensionality space). This technique has recently been used to develop automated classification of diseases and to improve methods for detecting disease in the clinical setting [ 4 , 5 ].

To test the potential power of SVM as an approach for classifying individuals into groups defined by disease status, we chose diabetes as an example. In the U.S., diabetes affects an estimated 23.6 million people, of whom about one third are unaware that they have the disease [ 6 ]. Another 57 million people have pre-diabetes, with elevated blood glucose levels that increase their risk of developing diabetes, heart disease, and stroke. Recent studies indicate that diabetes can be prevented by lifestyle changes or pharmacotherapy among persons with pre-diabetes [ 7 – 9 ]. Early screening and diagnosis is thus central to effective prevention strategies [ 10 ]. To this end, numerous risk scores and prediction equations have been developed to identify people at high risk of developing diabetes or with pre-diabetes based on common risk factors such as body mass index (BMI) and family history of diabetes [ 11 – 13 ]. For example, a recently published risk calculator uses logistic regression to identify people with pre-diabetes and undiagnosed diabetes by employing combinations of common risk variables [ 14 ]. Our objective was to generate an SVM-based approach to distinguish people with either undiagnosed diabetes or pre-diabetes from people without either of these conditions. The variables used to generate the SVM models were limited to simple clinical measurements that do not require laboratory tests. Predictions from this approach were compared with the predictions from logistic regression models containing the same set of variables. A final goal was to demonstrate the applicability of the SVM approach by creating a demonstration web-based classification tool.

Data source

In this study, we used a 1999-2004 data set from the National Health and Nutrition Examination Survey (NHANES) to generate the SVM algorithm. NHANES is an ongoing, cross-sectional, probability sample survey of the U.S. population. It collects demographic, health history, and behavioral information from participants in home interviews. Participants are also invited for detailed physical, physiological, and laboratory examinations that are performed by trained personnel in specially equipped mobile centers [ 15 ].

We limited our study to non-pregnant participants aged 20 or older. Participants were considered to have diagnosed diabetes if they answered "yes" to the question "Have you ever been told by a doctor or health professionals that you have diabetes?" Participants who answered "no" to this question but who had a measured fasting plasma glucose ≥ 126 mg/dl were considered to have undiagnosed diabetes; those with a fasting plasma glucose 100-125 mg/dl were considered to have pre-diabetes. Participants with fasting glucose <100 mg/dl were considered to not have diabetes (Table 1 ).

We devised two different classification schemes (Table 1 ). In Classification Scheme I, the group of persons with diabetes (diagnosed or undiagnosed) was distinguished from those without diabetes, including persons with pre-diabetes. In Classification Scheme II, the group of persons with either undiagnosed diabetes or pre-diabetes was distinguished from those without diabetes. The models were developed using a sample of 80% of the individuals in each group and validated in the remaining 20%.

Variable selection

We selected 14 simple variables commonly associated with the risk for diabetes: family history, age, gender, race and ethnicity, weight, height, waist circumference, BMI, hypertension, physical activity, smoking, alcohol use, education, and household income. Variable selection was performed according to an automatic approach developed by Chen et al. [ 16 ]. The significance of the automatically selected set of variables was further manually evaluated by fine tuning parameters. The variables included in the final selection were those with the best discriminative performance.

Model generation

Support Vector Machine (SVM) is a supervised machine learning technique that is widely used in pattern recognition and classification problems. The SVM algorithm performs a classification by constructing a multidimensional hyperplane that optimally discriminates between two classes by maximizing the margin between two data clusters. This algorithm achieves high discriminative power by using special nonlinear functions called kernels to transform the input space into a multidimensional space [ 17 ].

The basic idea behind the SVM technique is to construct an n-1 dimensional separating hyperplane to discriminate two classes in an n-dimensional space. A data point is viewed as an n-dimensional vector. For example, two variables in a dataset will create a two-dimensional space; the separating hyperplane would be a straight line (one dimensional) dividing the space in half. When more dimensions are involved, SVM searches for an optimal separating hyperplane called the maximum-margin separating hyperplane. The distance between the hyperplane and the nearest data point on each side (called support vectors) is maximized. The best scenario is that two classes are separated by a linear hyperplane. However, real-world situations are not always that simple. Some data points in the two classes might fall into a "grey" area that is not easy to be separated. SVM solves this problem by 1) allowing some data points to the wrong side of the hyperplane by introducing a user-specified parameter C that specifies the trade-off between the minimization of the misclassifications and maximization of margin; 2) using kernel functions (usually including linear, polynomial, sigmoid, and radial basis functions (RBF)) to add more dimensions to the low dimensional space, as a result that two classes could be separable in the high dimensional space. Figure 1 shows an example of an inseparable two-dimensional space that becomes separable after the transformation of the input space from low dimensional to multi dimensional. The SVM approach tends to classify entities without providing estimates of the probabilities of class membership in the dataset, which is a fundamental difference from multiple logistic regression.

Demonstration of finding a separating hyperplane in high dimensional space vs in low dimensional space .

Two key parameters for the kernels, C and gamma, need to be pre-selected to generate an optimal SVM model. Parameter C controls over-fitting of the model by specifying tolerance for misclassification. Parameter gamma controls the degree of nonlinearity of the model.

We used LibSVM [ 18 ], a freely available SVM software library, to generate the SVM models. To generate the data set for model training, we randomly selected a number of non-cases to match the number of cases in the training data set (see Table 1 for the definitions of cases and non-cases). According to the required data format input, values of selected features were normalized to values from - 1 to +1. Values of categorical variables such as Race are arbitrarily assigned to numbers between -1 and +1. For example, -1, -0.5, 0, 0.5, 1 represents non-Hispanic white, non-Hispanic black, Mexican American, other, other Hispanic respectively. Values of continuous variables were transformed into values between -1 and +1 by dividing them by an appropriate number. For example, the age values were divided by 100. In the training data set, the first column of the input data was set to the known outcome, i.e., 1 for positive, - 1 for negative. A utility included in the LibSVM package (grid.py) was used to find the optimal parameters for penalty parameter C and gamma under 5-fold cross-validation. Different kernel functions, including linear, polynomial, sigmoid, and radial basis functions (RBF), were tested and selected for the models based on performance.

Multiple logistic regression modeling (MLR) was performed using the same selected risk variables or features and case status (as specified previously and in Table 1 ) as the outcome variable. The logistic regression analysis was performed with the training data set using SAS-callable SUDAAN version 9, a procedure specific for complex survey design. Then, the estimated β coefficients were applied to the test data set to calculate for each individual the probability of being a case.

Model evaluations

Evaluation in the test data sets.

Test data sets were used to assess the performance of the models. Validation using the test data sets avoided potential bias of the performance estimate due to over-fitting of the model to training data sets. For the SVM model, the data files in the test data sets were formatted according to the requirement that variable values be normalized to values from - 1 to +1; the first column of the input data set (indicating case status) was set to 0. Prediction program Java code from the LibSVM library was modified to output the decision value (internal score generated by SVM tool) for each member of the test data set. For the logistic regression model, the prediction value for each member of the test data set was estimated by using the logistic regression function generated during the training step.

10-fold cross-validation in the training data set

To evaluate the robustness of the estimates from the SVM models, a 10-fold cross-validation was performed in the training data set. The training data set was partitioned into 10 equal-size subsets. Each subset was used as a test data set for a model trained on all cases and an equal number of non-cases randomly selected from the 9 remaining data subsets. This cross-validation process was repeated 10 times, allowing each subset to serve once as the test data set. To generate summary performance estimates, we averaged the area under the curve (AUC) of the receiver operating characteristic (ROC) curve and other statistics (sensitivity, specificity, positive predictive value [PPV], negative predictive value [NPV]) of the cross-validations.

Statistics for performance evaluation

ROC curves were generated based on the predicted outcome and true outcome. The AUCs for the test data sets were calculated and used to compare the discriminative powers of the models. We used Delong's method to calculate P -values to compare the AUCs based on results of the SVM models and MLR models [ 19 ].

Sensitivity, specificity, PPV, and NPV were calculated based on the following formulas when the cutoff value was set to default value (0) in the SVM model.

where TP, FP, TN, and FN represent the number of true positives, false positives, true negatives, and false negatives, respectively.

Demonstration web-based classification tool implementation

We implemented the SVM model as a web-based tool that we called Diabetes Classifier. The application was built by using J2EE technology [ 20 ] and other Java open-source frameworks such as Hibernate [ 21 ] and Strut [ 22 ]. LibSVM open-source Java codes were modified and embedded in the system source codes for prediction. The lookup tables for cutoff values and corresponding statistics (sensitivity, specificity) were generated from the calculations on each data point in the test data sets. Diabetes Classifier is freely accessible via http://www.hugenavigator.net/DiseaseClassificationPortal/startPageDiabetes.do

Results and Discussion

In Classification Scheme I (diagnosed or undiagnosed diabetes vs. no diabetes or pre-diabetes), 8 variables--family history, age, race and ethnicity, weight, height, waist circumference, BMI, and hypertension--yielded the best performance. In Classification Scheme II (undiagnosed diabetes or pre-diabetes vs. no diabetes), 10 variables--family history, age, race and ethnicity, weight, height, waist circumference, BMI, hypertension, sex, and physical activity--performed best. Kernel functions were evaluated in terms of their discriminative accuracy by AUC. The RBF kernel function performed best in Classification Scheme I, and the linear kernel function performed best in Classification Scheme II (Table 2 ). Performance parameters such as the AUC, sensitivity, specificity, positive predictive value, and negative predictive value are presented in Table 3 . The overall discriminative ability of Classification Schemes I and II are represented by their AUC values (83.47% and 73.18%, respectively; Figure 2 ).

ROC curves for Classifications Schemes I (a) and II (b) with SVM models and logistic regression models . Note: see Table 1 for the definitions of Classification Schemes I and II.

The AUC values for logistic regression analyses of the Classification Schemes I and II were 83.19% and 73.35%, respectively (Figure 2 ). Comparing the AUCs from our SVM and MLR models revealed no statistically significant difference in their discriminative abilities ( P = 0.3672 and P = 0.6718 for Classification Schemes I and II, respectively); thus, the SVM approach appears to perform as well as the traditional logistic regression model.

Diabetes Classifier, the web-based demonstration tool, was built based on the two SVM models. By selecting one of the diabetes classifications, the user is asked to enter the values for 8 or 10 common variables; the classification result is then presented on the next page, using the default cutoff value (0). This application provides an interface that allows the user to select the cutoff values. Each cutoff value has specific values of sensitivity and specificity so that the user can decide how the tool could be used in screening for diabetes. Diabetes Classifier can also be used in batch mode to classify observations in an uploaded file containing appropriately formatted values of required variables.

In this study, we tested two classification schemes to detect cases of diabetes and pre-diabetes in the U.S. population. Both schemes are examples of the potential use of support vector machine techniques in the classification of common diseases. Our results demonstrated that the discriminative performance of SVM models was equivalent to the epidemiological method commonly used for this purpose, multivariate logistic regression. To our knowledge, this is the first report that the SVM approach can be used successfully to detect a common disease with simple clinical measurements, without laboratory tests. Based on these results, we also developed a web-based tool for classification of diabetes and pre-diabetes. This tool demonstrates useful features for the potential application of classification algorithms in health care.

SVM is a model-free method that provides efficient solutions to classification problems without any assumption regarding the distribution and interdependency of the data. In epidemiologic studies and population health surveys, the SVM technique has the potential to perform better than traditional statistical methods like logistic regression, especially in situations that include multivariate risk factors with small effects (e.g., genome-wide association data and gene expression profiles), limited sample size, and a limited knowledge of underlying biological relationships among risk factors. This is particularly true in the case of common complex diseases where many risk factors, including gene-gene interactions and gene-environment interactions, have to be considered to reach sufficient discriminative power in prediction models [ 23 ]. Our work provides a promising proof of principle by demonstrating the predictive power of the SVM with just a small set of variables. This approach can be extended to include large data sets, including many other variables, such as genetic biomarkers, as data become available.

A major strength of this study is that we used the NHANES data set, which is a unique national weighted survey data that is representative of the U.S. population. Our results are comparable to those of other models tested in the same population. For example, Keikes et al. [ 24 ] developed a tool for detecting undiagnosed diabetes and pre-diabetes using logistic regression and a classification tree method to predict the risk of the diabetes in the U.S. population. Although direct comparisons are difficult because of the use of different NHANES data sets and different validation strategies, the discriminative powers in both studies seem to be equivalent. In our study, the AUC for the detection of diagnosed diabetes or undiagnosed diabetes was 83.47%, and it was 73.18% for pre-diabetes or undiagnosed diabetes in the validation test. In the study from Keikes et al., the AUC for undiagnosed diabetes were 82.19% (5-fold cross-validation) and 75.03% (training data set) for pre-diabetes or undiagnosed diabetes. Schwarz et al. [ 25 ] recently published a comprehensive review of existing tools for predicting the risk of type 2 diabetes or detecting undiagnosed diabetes. These tools were developed for different populations under different methodologies using different sets of variables. In general, the discriminative power of our SVM method is within the range of discriminative powers reported for the tools included in this review.

We cannot be certain that the models we developed by using the particular NHANES data set described here are applicable to other populations. Our SVM approach, however, is easily extended to other populations to generate their own classification systems. Likewise, a similar approach could be used to develop SVM models for other complex diseases using a different set of relevant variables.

A critical step for determining the usefulness of a screening test is to establish optimal cutoff values that yield optimal sensitivity and specificity values, which are particularly important for cost-effectiveness analysis [ 26 ]. Our web-based application, Diabetes Classifier, displays the trade-offs in sensitivity and specificity of the classification method as the cutoff value is changed. This feature is particularly relevant to clinical and public health programs, which can configure cutoff scores according to the objectives of the program and other considerations including cost-effectiveness. Diabetes Classifier allows data to be fed automatically (via data batch file uploading) for classification and provides an interface capable of sharing information with other sectors of a health care system. Web-based tools such as Diabetes Classifier can also serve as self-assessment tools for use by the general public.

Support vector machine modeling is a promising classification approach for detecting a complex disease like diabetes using common, simple variables. Validation indicated that the discriminative powers of our two SVM models are comparable to those of commonly used multivariable logistic regression methods. Our Diabetes Classifier tool, a web-based tool developed for demonstration purposes only, illustrates a potential use of the SVM technique: the identification of people with undetected common diseases such as diabetes and pre-diabetes. This approach needs to tested and validated in other studies.

Cortes C, Vapnik V: Support-vector networks. Machine Learning. 1995, 20: 273-297.

Google Scholar  

Ng KL, Mishra SK: De novo SVM classification of precursor microRNAs from genomic pseudo hairpins using global and intrinsic folding measures. Bioinformatics. 2007, 23: 1321-1330. 10.1093/bioinformatics/btm026.

Article   CAS   PubMed   Google Scholar  

Rice SB, Nenadic G, Stapley BJ: Mining protein function from text using term-based support vector machines. BMC Bioinformatics. 2005, 6 (Suppl 1): S22-10.1186/1471-2105-6-S1-S22.

Article   PubMed   PubMed Central   Google Scholar  

Maglogiannis I, Loukis E, Zafiropoulos E, Stasis A: Support Vectors Machine-based identification of heart valve diseases using heart sounds. Comput Methods Programs Biomed. 2009, 95: 47-61. 10.1016/j.cmpb.2009.01.003.

Article   PubMed   Google Scholar  

Thurston RC, Matthews KA, Hernandez J, De La TF: Improving the performance of physiologic hot flash measures with support vector machines. Psychophysiology. 2009, 46: 285-292. 10.1111/j.1469-8986.2008.00770.x.

American Diabetes Association. 2009, [ http://www.diabetes.org/ ]

Pi-Sunyer FX: How effective are lifestyle changes in the prevention of type 2 diabetes mellitus?. Nutr Rev. 2007, 65: 101-110. 10.1111/j.1753-4887.2007.tb00287.x.

Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM: Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002, 346: 393-403. 10.1056/NEJMoa012512.

Fruchter O: Prevention of type 2 diabetes mellitus by changes in lifestyle. N Engl J Med. 2001, 345: 696-697. 10.1056/NEJM200108303450912.

Global Guideline for Type 2 Diabetes: recommendations for standard, comprehensive, and minimal care. Diabet Med. 2006, 23: 579-593. 10.1111/j.1464-5491.2006.01918.x.

Thomas C, Hypponen E, Power C: Type 2 diabetes mellitus in midlife estimated from the Cambridge Risk Score and body mass index. Arch Intern Med. 2006, 166: 682-688. 10.1001/archinte.166.6.682.

Franciosi M, De BG, Rossi MC, Sacco M, Belfiglio M, Pellegrini F, Tognoni G, Valentini M, Nicolucci A: Use of the diabetes risk score for opportunistic screening of undiagnosed diabetes and impaired glucose tolerance: the IGLOO (Impaired Glucose Tolerance and Long-Term Outcomes Observational) study. Diabetes Care. 2005, 28: 1187-1194. 10.2337/diacare.28.5.1187.

Standards of medical care in diabetes--2008. Diabetes Care. 2008, 31 (Suppl 1): S12-S54.

Heikes KE, Eddy DM, Arondekar B, Schlessinger L: Diabetes Risk Calculator: a simple tool for detecting undiagnosed diabetes and pre-diabetes. Diabetes Care. 2008, 31: 1040-1045. 10.2337/dc07-1150.

National Health and Nutrition Examination Survey(NHANES). 2009, [ http://www.cdc.gov/nchs/nhanes/about_nhanes.htm ]

Yi-Wei Chen C-JL: Combining SVMs with Various Feature Selection Strategies. Feature Extraction: Foundations and Applications (Studies in Fuzziness and Soft Computing). 2005, Isabelle Guyon SGMNLAZ

Cover TC: Geometrical and statistical properties of system of linear inequalities with applications in pattern recognition [abstract]. IEEE Trans Electr Comput. 1965, 14: 326-10.1109/PGEC.1965.264137.

Article   Google Scholar  

Chang CC, Lin CJ: LIBSVM: a library for supportvector machines, 2001. 2007, [ http://www.csie.ntu.edu.tw/~cjlin/libsvm/ ]

DeLong ER, DeLong DM, Clarke-Pearson DL: Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988, 44: 837-845. 10.2307/2531595.

Java J2EE. Sun Microsystems, Inc. 2006, [ http://java.sun.com/javaee/ ]

Hibernate. JBoss Enterprise Middleware System. 2006, [ http://www.hibernate.org/ ]

Apache Struts. The Apache Software Foundation. 2006, [ http://struts.apache.org/ ]

Schadt EE, Friend SH, Shaywitz DA: A network view of disease and compound screening. Nat Rev Drug Discov. 2009, 8: 286-295. 10.1038/nrd2826.

Schwarz PE, Li J, Lindstrom J, Tuomilehto J: Tools for predicting the risk of type 2 diabetes in daily practice. Horm Metab Res. 2009, 41: 86-97. 10.1055/s-0028-1087203.

Cantor SB, Sun CC, Tortolero-Luna G, Richards-Kortum R, Follen M: A comparison of C/B ratios from studies using receiver operating characteristic curve analysis. J Clin Epidemiol. 1999, 52: 885-892. 10.1016/S0895-4356(99)00075-X.

Pre-publication history

The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1472-6947/10/16/prepub

Download references

Author information

Authors and affiliations.

National Office of Public Health Genomics, Coordinating Center for Health Promotion, Centers for Disease Control and Prevention, Atlanta, GA, USA

Wei Yu, Tiebin Liu, Rodolfo Valdez, Marta Gwinn & Muin J Khoury

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Wei Yu .

Additional information

Competing interests.

The authors declare that they have no competing interests.

Authors' contributions

WY designed and developed the methodology, built the demo web-based system, and drafted the manuscript. TL performed the data preparation and statistical analysis, RV provided expertise on diabetes and helped in manuscript preparation. MG provided advice on the project and revised the draft manuscript. MJK oversaw the project and revised the draft manuscript. All authors read and approved the final document.

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

Authors’ original file for figure 2, authors’ original file for figure 3, rights and permissions.

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article.

Yu, W., Liu, T., Valdez, R. et al. Application of support vector machine modeling for prediction of common diseases: the case of diabetes and pre-diabetes. BMC Med Inform Decis Mak 10 , 16 (2010). https://doi.org/10.1186/1472-6947-10-16

Download citation

Received : 18 December 2009

Accepted : 22 March 2010

Published : 22 March 2010

DOI : https://doi.org/10.1186/1472-6947-10-16

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Support Vector Machine
• Classification Scheme
• Support Vector Machine Model
• Undiagnosed Diabetes
• Support Vector Machine Algorithm

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

• General enquiries: [email protected]

IEEE Account

• Change Username/Password
• Update Address

Purchase Details

• Payment Options
• Order History
• View Purchased Documents

Profile Information

• Communications Preferences
• Profession and Education
• Technical Interests
• US & Canada: +1 800 678 4333
• Worldwide: +1 732 981 0060
• Contact & Support
• About IEEE Xplore
• Accessibility
• Terms of Use
• Nondiscrimination Policy
• Privacy & Opting Out of Cookies

A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. © Copyright 2024 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.

Stay up to date on the latest in Machine Learning and AI

Support vector machines.

Unlock the full potential of Support Vector Machines (SVMs) by learning how to implement Linear SVM, a crucial classification algorithm. In this article, we’ll delve into the theoretical foundations, …

Kernel Support Vector Machines (SVMs) in Python

In this article, we delve into the realm of non-linear classification using Kernel Support Vector Machines (SVMs), a powerful technique that’s essential for advanced machine learning projects. We’ll e …

Support Vector Machines for Regression

Explore the concept of Support Vector Machines (SVMs) for regression, a powerful machine learning algorithm used for predicting continuous outcomes. Dive into its theoretical foundations, practical ap …

Grab your spot at the free arXiv Accessibility Forum

Help | Advanced Search

Computer Science > Machine Learning

Title: quantum-inspired support vector machine.

Abstract: Support vector machine (SVM) is a particularly powerful and flexible supervised learning model that analyzes data for both classification and regression, whose usual algorithm complexity scales polynomially with the dimension of data space and the number of data points. To tackle the big data challenge, a quantum SVM algorithm was proposed, which is claimed to achieve exponential speedup for least squares SVM (LS-SVM). Here, inspired by the quantum SVM algorithm, we present a quantum-inspired classical algorithm for LS-SVM. In our approach, a improved fast sampling technique, namely indirect sampling, is proposed for sampling the kernel matrix and classifying. We first consider the LS-SVM with a linear kernel, and then discuss the generalization of our method to non-linear kernels. Theoretical analysis shows our algorithm can make classification with arbitrary success probability in logarithmic runtime of both the dimension of data space and the number of data points for low rank, low condition number and high dimensional data matrix, matching the runtime of the quantum SVM.

Submission history

Access paper:.

• Other Formats

References & Citations

• INSPIRE HEP
• Google Scholar
• Semantic Scholar

DBLP - CS Bibliography

Bibtex formatted citation.

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Development of a predictive model for carbon dioxide corrosion rate and severity based on machine learning algorithms.

1. Introduction

2. overview of algorithms and evaluation metrics for models, 2.1. introduction to machine learning algorithms, 2.1.1. overview of the random forest algorithm, 2.1.2. introduction to xgboost algorithm principle, 2.2. evaluation of multiple model, 2.2.1. performance evaluation metrics for regression models, 2.2.2. evaluation metrics for classification performance, 3. data mining, 3.1. data preprocessing, 3.2. correlation analysis, 4. discussion, 4.1. ml models for carbon dioxide corrosion rate, 4.2. ml models for carbon dioxide corrosion severity, 5. conclusions, author contributions, institutional review board statement, informed consent statement, data availability statement, acknowledgments, conflicts of interest, abbreviations.

• Xhanari, K.; Wang, Y.; Yang, Z.; Finšgar, M. A Review of Recent Advances in the Inhibition of Sweet Corrosion. Chem. Rec. 2021 , 21 , 1845–1875. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Yevtushenko, O.; Bettge, D.; Bohraus, S.; Bäßler, R.; Pfennig, A.; Kranzmann, A. Corrosion Behavior of Steels for CO 2 Injection. Process Saf. Environ. Prot. 2014 , 92 , 108–118. [ Google Scholar ] [ CrossRef ]
• Nesic, S.; Postlethwaite, J.; Olsen, S. An Electrochemical Model for Prediction of Corrosion of Mild Steel in Aqueous Carbon Dioxide Solutions. Corrosion 1996 , 52 , 280–294. [ Google Scholar ] [ CrossRef ]
• Bin, H.J.; Yang, Y.; Brown, B.; Nesic, S. Electrochemical Investigation of Localized CO 2 Corrosion on Mild Steel ; OnePetro: Richardson, TX, USA, 2007. [ Google Scholar ]
• Pessu, F.; Barker, R.; Neville, A. Pitting and Uniform Corrosion of X65 Carbon Steel in Sour Corrosion Environments: The Influence of CO 2 , H 2 S, and Temperature. Corrosion 2017 , 73 , 1168–1183. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Zhu, G.Y.; Li, Y.Y.; Hou, B.S.; Zhang, Q.H.; Zhang, G.A. Corrosion Behavior of 13Cr Stainless Steel under Stress and Crevice in High Pressure CO 2 /O 2 Environment. J. Mater. Sci. Technol. 2021 , 88 , 79–89. [ Google Scholar ] [ CrossRef ]
• Senatore, E.V.; Taleb, W.; Owen, J.; Hua, Y.; Gomes, J.A.C.P.; Barker, R.; Neville, A. Evaluation of High Shear Inhibitor Performance in CO 2 -Containing Flow-Induced Corrosion and Erosion-Corrosion Environments in the Presence and Absence of Iron Carbonate Films. Wear 2018 , 404–405 , 143–152. [ Google Scholar ] [ CrossRef ]
• Kahyarian, A.; Achour, M.; Nesic, S. Mathematical Modeling of Uniform CO 2 Corrosion. In Trends in Oil and Gas Corrosion Research and Technologies ; El-Sherik, A.M., Ed.; Woodhead Publishing Series in Energy; Woodhead Publishing: Boston, MA, USA, 2017; pp. 805–849. ISBN 978-0-08-101105-8. [ Google Scholar ]
• Nesic, S.; Nordsveen, M.; Maxwell, N.; Vrhovac, M. Probabilistic Modelling of CO 2 Corrosion Laboratory Data Using Neural Networks. Corros. Sci. 2001 , 43 , 1373–1392. [ Google Scholar ] [ CrossRef ]
• Jones, M.; Owen, J.; de Boer, G.; Woollam, R.C.; Folena, M.C.; Farhat, H.; Barker, R. Numerical Exploration of a Fully Mechanistic Mathematical Model of Aqueous CO 2 Corrosion in Steel Pipelines. Corros. Sci. 2024 , 236 , 112235. [ Google Scholar ] [ CrossRef ]
• Dana, M.M.; Javidi, M. Corrosion Simulation via Coupling Computational Fluid Dynamics and NORSOK CO 2 Corrosion Rate Prediction Model for an Outlet Header Piping of an Air-Cooled Heat Exchanger. Eng. Fail. Anal. 2021 , 122 , 105285. [ Google Scholar ] [ CrossRef ]
• De Waard, C.; Milliams, D.E. Carbonic Acid Corrosion of Steel. Corrosion 2013 , 31 , 177–181. [ Google Scholar ] [ CrossRef ]
• De Waard, C.; Lotz, U.; Milliams, D.E. Predictive Model for CO 2 Corrosion Engineering in Wet Natural Gas Pipelines. Corrosion 1991 , 47 , 976–985. [ Google Scholar ] [ CrossRef ]
• Waard, C.D.; Lotz, U. Prediction of CO 2 Corrosion of Carbon Steel ; Institute of Materials: London, UK, 1994. [ Google Scholar ]
• Eliyan, F.F.; Alfantazi, A. On the Theory of CO 2 Corrosion Reactions—Investigating Their Interrelation with the Corrosion Products and API-X100 Steel Microstructure. Corros. Sci. 2014 , 85 , 380–393. [ Google Scholar ] [ CrossRef ]
• Nesic, S.; Lee, J.; Ruzic, V. A Mechanistic Model of Iron Carbonate Film Growth and the Effect on CO 2 Corrosion of Mild Steel. In Proceedings of the CORROSION 2002, Denver, CO, USA, 7–11 April 2002. [ Google Scholar ]
• Abbas, M.H.; Norman, R.; Charles, A. Neural Network Modelling of High Pressure CO 2 Corrosion in Pipeline Steels. Process Saf. Environ. Prot. 2018 , 119 , 36–45. [ Google Scholar ] [ CrossRef ]
• Ossai, C.I. A Data-Driven Machine Learning Approach for Corrosion Risk Assessment—A Comparative Study. Big Data Cogn. Comput. 2019 , 3 , 28. [ Google Scholar ] [ CrossRef ]
• Yan, L.; Diao, Y.; Lang, Z.; Gao, K. Corrosion Rate Prediction and Influencing Factors Evaluation of Low-Alloy Steels in Marine Atmosphere Using Machine Learning Approach. Sci. Technol. Adv. Mater. 2020 , 21 , 359–370. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Aghaaminiha, M.; Mehrani, R.; Colahan, M.; Brown, B.; Singer, M.; Nesic, S.; Vargas, S.M.; Sharma, S. Machine Learning Modeling of Time-Dependent Corrosion Rates of Carbon Steel in Presence of Corrosion Inhibitors. Corros. Sci. 2021 , 193 , 109904. [ Google Scholar ] [ CrossRef ]
• Paul, A.; Mukherjee, D.P.; Das, P.; Gangopadhyay, A.; Chintha, A.R.; Kundu, S. Improved Random Forest for Classification. IEEE Trans. Image Process. 2018 , 27 , 4012–4024. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Dasarathy, B.V.; Sheela, B.V. A Composite Classifier System Design: Concepts and Methodology. Proc. IEEE 1979 , 67 , 708–713. [ Google Scholar ] [ CrossRef ]
• Chen, T.; Guestrin, C. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Fracisco, CA, USA, 13–17 August 2016; Association for Computing Machinery: New York, NY, USA, 2016; pp. 785–794. [ Google Scholar ]
• Farelas, F.; Choi, Y.S.; Nešić, S. Corrosion Behavior of API 5L X65 Carbon Steel Under Supercritical and Liquid Carbon Dioxide Phases in the Presence of Water and Sulfur Dioxide. Corrosion 2013 , 69 , 243–250. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Xiang, Y.; Wang, Z.; Li, Z.; Ni, W.D. Effect of Exposure Time on the Corrosion Rates of X70 Steel in Supercritical CO 2 /SO 2 /O 2 /H 2 O Environments. Corrosion 2013 , 69 , 251–258. [ Google Scholar ] [ CrossRef ]
• Wang, T.; Li, J.; Liu, Y. Based on Pearson Correlation Coefficient and Monte Carlo Simulation Method, the Calculation Formula of Frictional Resistance Is Optimized for Soft Shaft Bending and Pulling of Steel Wire. J. Phys. Conf. Ser. 2021 , 1939 , 012088. [ Google Scholar ] [ CrossRef ]

Click here to enlarge figure

Share and Cite

Dong, Z.; Zhang, M.; Li, W.; Wen, F.; Dong, G.; Zou, L.; Zhang, Y. Development of a Predictive Model for Carbon Dioxide Corrosion Rate and Severity Based on Machine Learning Algorithms. Materials 2024 , 17 , 4046. https://doi.org/10.3390/ma17164046

Dong Z, Zhang M, Li W, Wen F, Dong G, Zou L, Zhang Y. Development of a Predictive Model for Carbon Dioxide Corrosion Rate and Severity Based on Machine Learning Algorithms. Materials . 2024; 17(16):4046. https://doi.org/10.3390/ma17164046

Dong, Zhenzhen, Min Zhang, Weirong Li, Fenggang Wen, Guoqing Dong, Lu Zou, and Yongqiang Zhang. 2024. "Development of a Predictive Model for Carbon Dioxide Corrosion Rate and Severity Based on Machine Learning Algorithms" Materials 17, no. 16: 4046. https://doi.org/10.3390/ma17164046

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals

Wind power forecasting using a GRU attention model for efficient energy management systems

• Original Paper
• Published: 15 August 2024

Cite this article

• Lakhdar Nadjib Boucetta 1 ,
• Youssouf Amrane 1 &
• Saliha Arezki 1  

Modern energy management systems play a crucial role in integrating multiple renewable energy sources into electricity grids, enabling a balanced supply–demand relationship while promoting eco-friendly energy consumption. Among these renewables, wind energy, with its environmental and economic advantages, poses challenges due to its inherent variability, demanding accurate prediction models for seamless integration. This paper presents an innovative hybrid deep learning model that integrates a gated recurrent unit (GRU)-based attention mechanism neural network for wind power generation forecast. The developed model’s performance is compared against six other models, comprising four deep learning approaches—long short-term memory (LSTM), 1D convolutional neural network, convolutional neural short-term memory (CNN-LSTM), and convolutional long short-term memory (ConvLSTM)—as well as two machine learning models—random forest and support vector regression. The proposed GRU-based attention model demonstrates superior performance, particularly in 1-step to 3-step ahead predictions, with mean absolute error values of 59.45, 114.95, and 176.06, root mean square error values of 109.03, 201.83, and 296.55, normalized root mean square error values of 0.080, 0.148, and 0.218, and coefficient of determination (R2) values of 0.992, 0.975, and 0.948, for forecast horizons of 10, 20, and 30 min, respectively. These results underscore the robust predictive capability of the proposed algorithm. Significantly, this research constitutes the first application of the hybrid GRU-based attention model to the Yalova wind turbine dataset, achieving better accuracy when compared to prior studies utilizing the same data.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Explore related subjects

• Artificial Intelligence

Meliani M, Barkany AE, Abbassi IE et al (2021) Energy management in the smart grid: state-of-the-art and future trends. Int J Eng Bus Manag 13:18479790211032920. https://doi.org/10.1177/18479790211032920

Article   Google Scholar  

Wang H, Lei Z, Zhang X, Zhou B, Peng J (2019) A review of deep learning for renewable energy forecasting. Energy Convers Manag 198:111799. https://doi.org/10.1016/j.enconman.2019.111799

Hu J, Heng J, Wen J, Zhao W (2020) Deterministic and probabilistic wind speed forecasting with de-noising-reconstruction strategy and quantile regression-based algorithm. Renew Energy 162:1208–1226. https://doi.org/10.1016/j.renene.2020.08.077

Han Q, Meng F, Hu T, Chu F (2017) Non-parametric hybrid models for wind speed forecasting. Energy Convers Manag 148:554–568. https://doi.org/10.1016/j.enconman.2017.06.021

Yunus K, Thiringer T, Chen P (2016) ARIMA-based frequency-decomposed modeling of wind speed time series. IEEE Trans Power Syst 31:2546–2556. https://doi.org/10.1109/TPWRS.2015.2468586

Ait Maatallah O, Achuthan A, Janoyan K, Marzocca P (2015) Recursive wind speed forecasting based on Hammerstein Auto-Regressive model. Appl Energy 145:191–197. https://doi.org/10.1016/j.apenergy.2015.02.032

Demolli H, Dokuz AS, Ecemis A, Gokcek M (2019) Wind power forecasting based on daily wind speed data using machine learning algorithms. Energy Convers Manag 198:111823. https://doi.org/10.1016/j.enconman.2019.111823

Li L-L, Zhao X, Tseng M-L, Tan RR (2020) Short-term wind power forecasting based on support vector machine with improved dragonfly algorithm. J Clean Prod 242:118447. https://doi.org/10.1016/j.jclepro.2019.118447

Ding M, Zhou H, Xie H et al (2021) A time series model based on hybrid-kernel least-squares support vector machine for short-term wind power forecasting. ISA Trans 108:58–68. https://doi.org/10.1016/j.isatra.2020.09.002

Garg S, Krishnamurthi R (2023) A CNN encoder decoder LSTM model for sustainable wind power predictive analytics. Sustain Comput Inf Syst 38:100869. https://doi.org/10.1016/j.suscom.2023.100869

Huang J, Niu G, Guan H, Song S (2023) Ultra-short-term wind power prediction based on LSTM with loss shrinkage adam. Energies 16:3789. https://doi.org/10.3390/en16093789

Shahid F, Zameer A, Muneeb M (2021) A novel genetic LSTM model for wind power forecast. Energy 223:120069. https://doi.org/10.1016/j.energy.2021.120069

Zhao Z, Yun S, Jia L et al (2023) Hybrid VMD-CNN-GRU-based model for short-term forecasting of wind power considering spatio-temporal features. Eng Appl Artif Intell 121:105982. https://doi.org/10.1016/j.engappai.2023.105982

Yildiz C, Acikgoz H, Korkmaz D, Budak U (2021) An improved residual-based convolutional neural network for very short-term wind power forecasting. Energy Convers Manag 228:113731. https://doi.org/10.1016/j.enconman.2020.113731

Sun Z, Zhao M (2020) Short-term wind power forecasting based on VMD decomposition, ConvLSTM networks and error analysis. IEEE Access 8:134422–134434. https://doi.org/10.1109/ACCESS.2020.3011060

Xiang L, Liu J, Yang X et al (2022) Ultra-short term wind power prediction applying a novel model named SATCN-LSTM. Energy Convers Manag 252:115036. https://doi.org/10.1016/j.enconman.2021.115036

Xiong B, Lou L, Meng X et al (2022) Short-term wind power forecasting based on attention mechanism and deep learning. Electr Power Syst Res 206:107776. https://doi.org/10.1016/j.epsr.2022.107776

Bentsen LØ, Warakagoda ND, Stenbro R, Engelstad P (2023) Spatio-temporal wind speed forecasting using graph networks and novel Transformer architectures. Appl Energy 333:120565. https://doi.org/10.1016/j.apenergy.2022.120565

Yu G, Liu C, Tang B et al (2022) Short term wind power prediction for regional wind farms based on spatial-temporal characteristic distribution. Renew Energy 199:599–612. https://doi.org/10.1016/j.renene.2022.08.142

Zhang J, Liu D, Li Z et al (2021) Power prediction of a wind farm cluster based on spatiotemporal correlations. Appl Energy 302:117568. https://doi.org/10.1016/j.apenergy.2021.117568

Gong M, Li W, Yan C, Liu Y, Li S, Zhao Z, Xu W (2023) Wind power forecasting based on SCINet, reversible instance normalization, and knowledge distillation. J Renew Sustain Energy 15:053306. https://doi.org/10.1063/5.0166061

Abou Houran M, Salman Bukhari SM, Zafar MH et al (2023) COA-CNN-LSTM: Coati optimization algorithm-based hybrid deep learning model for PV/wind power forecasting in smart grid applications. Appl Energy 349:121638. https://doi.org/10.1016/j.apenergy.2023.121638

Zheng J, Du J, Wang B et al (2023) A hybrid framework for forecasting power generation of multiple renewable energy sources. Renew Sustain Energy Rev 172:113046. https://doi.org/10.1016/j.rser.2022.113046

Wang H-K, Song K, Cheng Y (2022) A hybrid forecasting model based on CNN and informer for short-term wind power. Front Energy Res. https://doi.org/10.3389/fenrg.2021.788320

Chung J, Gulcehre C, Cho K, Bengio Y (2014) Empirical evaluation of gated recurrent neural networks on sequence modeling. In: arXiv.org. https://arxiv.org/abs/1412.3555v1

Boucetta LN, Amrane Y, Arezki S (2023) Comparative analysis of LSTM, GRU, and MLP neural networks for short-term solar power forecasting. In: 2023 international conference on electrical engineering and advanced technology (ICEEAT), pp 1–6

Niu Z, Zhong G, Yu H (2021) A review on the Attention mechanism of deep learning. Neurocomputing 452:48–62. https://doi.org/10.1016/j.neucom.2021.03.091

Guo M-H, Xu T-X, Liu J-J, Liu Z-N, Jiang P-T, Mu T-J, Zhang S-H, Martin RR, Cheng M-M, Hu S-M (2022) Attention mechanisms in computer vision: a survey. Comput Visual Media 8:331–368. https://doi.org/10.1007/s41095-022-0271-y

Bahdanau D, Cho K, Bengio Y (2014) Neural machine translation by jointly learning to align and translate. In: arXiv.org. https://arxiv.org/abs/1409.0473v7

Erisen, B. Wind Turbine Scada Dataset (2018) Available online: http:// www.kaggle.com/berkerisen/wind-turbine-scada-dataset . Accessed on 23 April 2023

Rahimi N, Park S, Choi W et al (2023) A comprehensive review on ensemble solar power forecasting algorithms. J Electr Eng Technol 18:719–733. https://doi.org/10.1007/s42835-023-01378-2

Chicco D, Warrens MJ, Jurman G (2021) The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput Sci 7:e623. https://doi.org/10.7717/peerj-cs.623

Zhao Y, Jia L (2020) A short-term hybrid wind power prediction model based on singular spectrum analysis and temporal convolutional networks. J Renew Sustain Energy 12:056101. https://doi.org/10.1063/5.0007003

Delgado I, Fahim M (2021) Wind turbine data analysis and LSTM-based prediction in SCADA system. Energies 14:125. https://doi.org/10.3390/en14010125

Wei H, Wang W, Kao X (2023) A novel approach to ultra-short-term wind power prediction based on feature engineering and informer. Energy Rep 9:1236–1250. https://doi.org/10.1016/j.egyr.2022.12.062

Download references

Author information

Authors and affiliations.

LSEI Laboratory, Department of Electrical Engineering, Faculty of Electronics and Computer Science, University of Science and Technology Houari Boumediene, Algiers, Algeria

Lakhdar Nadjib Boucetta, Youssouf Amrane & Saliha Arezki

You can also search for this author in PubMed   Google Scholar

Contributions

BOUCETTA Lakhdar Nadjib carried out the conceptualization, methodology, and original draft preparation and writing. Youssouf Amrane and Saliha Arezki participated with BOUCETTA Lakhdar Nadjib in investigation, the three authors collaborated on editing, and visualization. All authors reviewed the manuscript.

Corresponding author

Correspondence to Lakhdar Nadjib Boucetta .

Ethics declarations

Conflict of interest.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Boucetta, L.N., Amrane, Y. & Arezki, S. Wind power forecasting using a GRU attention model for efficient energy management systems. Electr Eng (2024). https://doi.org/10.1007/s00202-024-02590-7

Download citation

Received : 20 January 2024

Accepted : 01 July 2024

Published : 15 August 2024

DOI : https://doi.org/10.1007/s00202-024-02590-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Wind energy
• Energy management system (EMS)
• Wind power forecasting
• Deep learning
• GRU-based attention mechanism
• Find a journal
• Publish with us
• Track your research