Developing new VO 2 max prediction models from maximal, submaximal and questionnaire variables using support vector machines combined with feature selection

Abut, Fatih; Akay, Mehmet Fatih; George, James D.

doi:10.1016/j.compbiomed.2016.10.018

Cited by 30 publications

(30 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Particularly, compared with the RMSEs obtained by GRNN-based and SDT-based models, the average percentage decrement rates in RMSEs obtained by SVM-based models are 13.63% and 35.50% for prediction of VO 2 max, respectively. These results complement the previous studies in [3][4][5], where SVM was also reported to be superior to MLP, RBFNN, DTF, TB, and MLR for prediction of VO 2 max. GRNN-based models, in turn, outperform the SDT-based models, which comparatively show the worst performance for prediction of VO 2 max.…”

Section: Discussionsupporting

confidence: 90%

“…Secondly, MVFS is made up of Relief-F, mRMR, and MLFS, which are currently not included in EFS. Differently from the rest of the individual feature selectors implemented in EFS, Relief-F, mRMR, and MLFS have the merit that in previous studies [4,5,26] they have, in general, been shown to perform with satisfactory performance in creating accurate VO 2 max models. Finally, in EFS there is no weighting among the feature selectors.…”

Section: Results For Comparing Mvfs With Ensemble Feature Selectormentioning

confidence: 99%

“…Table 7 and Table 8 show the ranges of values of the utilized parameters for the SVM-based models and the GRNNand SDT-based models, respectively. [3,10] Maximum tree levels [5,40] The prediction errors of the models have been assessed by calculating the Rs and RMSEs, whose equations are given in Eq. (4) and Eq.…”

Section: Modelsmentioning

confidence: 99%

“…For comparison purposes, prediction models based on a radial basis function neural network (RBFNN) and decision tree forest (DTF) have also been created. In a follow-up work, the study in [5] proposed new hybrid VO 2 max models by combining the maximal, submaximal, and questionnaire variables using SVM combined with the Relief-F algorithm and compared the results to those of MLP-based and tree boost (TB)-based models. To sum up the major conclusions made by these three studies, it has been reported that sex, age, MX-HR, maximal exercise time, submaximal ending speed of the treadmill, and Perceived Functional Ability (Q-PFA) and Activity Code (AC) questionnaires are the essential predictors of VO 2 max, and SVM has been reported to be superior to the rest of the utilized regression methods for prediction of VO 2 max.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A robust ensemble feature selector based on rank aggregation for developing new VO2max prediction models using support vector machines

Abut¹,

Akay²,

George³

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

This paper proposes a new ensemble feature selector, called the majority voting feature selector (MVFS), for developing new maximal oxygen uptake (VO2max) prediction models using a support vector machine (SVM). The approach is based on rank aggregation, which meaningfully utilizes the correlation among the relevance ranks of predictor variables given by three state-of-the-art feature selectors: Relief-F, minimum redundancy maximum relevance (mRMR), and maximum likelihood feature selection (MLFS). By applying the SVM combined with MVFS on a self-created dataset containing maximal and submaximal exercise data from 185 college students, several new hybrid VO2max prediction models have been created. To compare the performance of the proposed ensemble approach on prediction of VO2max, SVM-based models with individual combinations of Relief-F, mRMR, and MLFS as well as with other alternative ensemble feature selectors from the literature have also been developed. The results reveal that MVFS outperforms other individual and ensemble feature selectors and yields up to 8.76% increment and 11.15% decrement rates in multiple correlation coefficients (Rs) and root mean square errors (RMSEs), respectively. Furthermore, in addition to reconfirming the relevance of sex, age, and maximal heart rate in predicting VO2max, which were previously reported in the literature, it is revealed that submaximal heart rates and exercise times at 1.5-mile distance are two further discriminative predictors of VO2max. The results have also been compared to those obtained by a general regression neural network and single decision tree combined with MVFS, and it is shown that the SVM exhibits much better performance than other methods for prediction of VO2max.

show abstract

Section: Discussionsupporting

confidence: 90%

Section: Results For Comparing Mvfs With Ensemble Feature Selectormentioning

confidence: 99%

Section: Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A robust ensemble feature selector based on rank aggregation for developing new VO2max prediction models using support vector machines

Abut¹,

Akay²,

George³

2019

Turk J Elec Eng & Comp Sci

Self Cite

View full text Add to dashboard Cite

show abstract

“…VO2max plays a significant role in sport science, education and research. In sport sciences, it is often used as an indicator for endurance capacity of athletes, representing the upper limit of their aerobic fitness (Abut, Akay & George, 2016).…”

Section: Introductionmentioning

confidence: 99%

New prediction models for the maximal oxygen uptake of collegeaged students using non-exercise data

Akay

Çetin²,

Yarım³

et al. 2017

PROSOC

Self Cite

View full text Add to dashboard Cite

Maximal oxygen uptake (VO2max) refers to the maximal amount of oxygen that an individual can utilize during intense or maximal exercise. VO2max plays a significant role in sport science, education and research. The direct measurement of VO2max is time-consuming, requires expensive laboratory equipment and trained staff. Because of these disadvantages of direct measurement, numerous VO2max prediction models for a variety of subject groups have been developed. The purpose of this study is to develop new Multiple Linear Regression based on VO2max prediction models for Turkish college students by using physiological and questionnaire variables. The dataset includes the data of 62 (28 females and 34 males) students, ranging in age from 18 to 27 years, from the College of Physical Education and Sports Science at Gazi University. Seven different models consisting of the predictor variables gender, age, weight, height, Perceived Functional Ability scores (PFA-1 and PFA-2), and Physical Activity Rating score (PA-R) have been used to predict VO2max. The performance of the prediction models has been evaluated by calculating their standard error of estimates (SEE's) and multiple correlation coefficients (R's). The prediction model including Gender, Age, Height, Weight, PFA-1 and PAR yields the lowest SEE with 5.14 mL.kg-1.min-1 and highest R with 0.93. It can be concluded that in situations where it is difficult to measure VO2max, the given model with MLR equation can be used to predict the VO2max of college students with acceptable error rates.

show abstract

Predicting renewable energy production by machine learning methods: The case of Turkey

Yağmur

Kayakuş

Terzioğlu

2023

Env Prog and Sustain Energy

View full text Add to dashboard Cite

It is considered that the use of renewable energy sources will replace fossil fuels due to global climate change and accompanying decisions taken by states. In this study, unlike the renewable energy production estimation studies in the literature, a model was created by taking the socioeconomic, environmental and energy time series data of the countries. In the study, Turkey, which did not promise a numerical reduction in greenhouse gas emissions unlike other developing countries but has an increasing energy production from renewable energy sources, was chosen. In the study, the data between 1990 and 2020 were used to receive more realistic results by considering the interval before and after the Kyoto protocol. Artificial neural networks and support vector regression among machine learning methods were used to predict the model. As a result of the study, support vector regression had a 92% and artificial neural networks had a successful predictive power of 89.9% according to the coefficient of determination (R 2 ). In the study, the root mean square error value was 0.071 for artificial neural networks and 0.045 for support vector regression; the mean squared error value was 0.005 for artificial neural networks and 0.002 for support vector regression, which was close to the ideal values. Both methods were statistically successful. It is predicted that the model designed because of these successful results obtained in the study would guide the creation of energy policies and contribute to scientific studies.

show abstract

Developing new VO 2 max prediction models from maximal, submaximal and questionnaire variables using support vector machines combined with feature selection

Cited by 30 publications

References 31 publications

A robust ensemble feature selector based on rank aggregation for developing new VO2max prediction models using support vector machines

A robust ensemble feature selector based on rank aggregation for developing new VO2max prediction models using support vector machines

New prediction models for the maximal oxygen uptake of collegeaged students using non-exercise data

Predicting renewable energy production by machine learning methods: The case of Turkey

Contact Info

Product

Resources

About