Next Article in Journal
Solvent-Free Synthesis of Magnetic Sewage Sludge-Derived Biochar for Heavy Metal Removal from Wastewater
Next Article in Special Issue
Implementation of a Community-Based Mind–Body (Tae-Bo) Physical Activity Programme on Health-Related Physical Fitness in Rural Black Overweight and Obese Women with Manifest Risk Factors for Multimorbidity
Previous Article in Journal
An Empirical Study on the Promotion of Students’ Physiological and Psychological Recovery in Green Space on Campuses in the Post-Epidemic Era
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJERPH
Volume 20
Issue 1
10.3390/ijerph20010146
ijerph-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Profiling Physical Fitness of Physical Education Majors Using Unsupervised Machine Learning

by
Diego A. Bonilla
1,2,3,4,
Isabel A. Sánchez-Rojas
5,*,
Darío Mendoza-Romero
5,
Yurany Moreno
1,
Jana Kočí
1,6,
Luis M. Gómez-Miranda
7,
Daniel Rojas-Valverde
8,9,
Jorge L. Petro
1,3 and
Richard B. Kreider
10
1
Research Division, Dynamical Business & Science Society—DBSS International SAS, Bogotá 110311, Colombia
2
Research Group in Biochemistry and Molecular Biology, Universidad Distrital Francisco José de Caldas, Bogotá 110311, Colombia
3
Research Group in Physical Activity, Sports and Health Sciences (GICAFS), Universidad de Córdoba, Montería 230002, Colombia
4
Sport Genomics Research Group, Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Science and Technology, University of the Basque Country (UPV/EHU), 48940 Leioa, Spain
5
Grupo de Investigación Ciencias Aplicadas al Ejercicio, Deporte y Salud—GICAEDS, Universidad Santo Tomás, Bogotá 205070, Colombia
6
Department of Education, Faculty of Education, Charles University, 11636 Prague, Czech Republic
7
Sports Faculty, Autonomous University of Baja California, Tijuana 22390, Mexico
8
Núcleo de Estudios para el Alto Rendimiento y la Salud (NARS-CIDISAD), Escuela Ciencia del Movimiento Humano y Calidad de Vida (CIEMHCAVI), Universidad Nacional, Heredia 863000, Costa Rica
9
Clínica de Lesiones Deportivas (Rehab&Readapt), Escuela Ciencia del Movimiento Humano y Calidad de Vida (CIEMHCAVI), Universidad Nacional, Heredia 863000, Costa Rica
10
Exercise & Sport Nutrition Laboratory, Human Clinical Research Facility, Texas A&M University, College Station, TX 77843, USA
*
Author to whom correspondence should be addressed.
Int. J. Environ. Res. Public Health 2023, 20(1), 146; https://doi.org/10.3390/ijerph20010146
Submission received: 18 November 2022 / Revised: 17 December 2022 / Accepted: 20 December 2022 / Published: 22 December 2022
(This article belongs to the Special Issue Editorial Board Members' Collection Series: Physical Activity and Health)
Browse Figures
Review Reports Versions Notes

Abstract

:
The academic curriculum has shown to promote sedentary behavior in college students. This study aimed to profile the physical fitness of physical education majors using unsupervised machine learning and to identify the differences between sexes, academic years, socioeconomic strata, and the generated profiles. A total of 542 healthy and physically active students (445 males, 97 females; 19.8 [2.2] years; 66.0 [10.3] kg; 169.5 [7.8] cm) participated in this cross-sectional study. Their indirect VO2max (Cooper and Shuttle-Run 20 m tests), lower-limb power (horizontal jump), sprint (30 m), agility (shuttle run), and flexibility (sit-and-reach) were assessed. The participants were profiled using clustering algorithms after setting the optimal number of clusters through an internal validation using R packages. Non-parametric tests were used to identify the differences (p < 0.05). The higher percentage of the population were freshmen (51.4%) and middle-income (64.0%) students. Seniors and juniors showed a better physical fitness than first-year students. No significant differences were found between their socioeconomic strata (p > 0.05). Two profiles were identified using hierarchical clustering (Cluster 1 = 318 vs. Cluster 2 = 224). The matching analysis revealed that physical fitness explained the variation in the data, with Cluster 2 as a sex-independent and more physically fit group. All variables differed significantly between the sexes (except the body mass index [p = 0.218]) and the generated profiles (except stature [p = 0.559] and flexibility [p = 0.115]). A multidimensional analysis showed that the body mass, cardiorespiratory fitness, and agility contributed the most to the data variation so that they can be used as profiling variables. This profiling method accurately identified the relevant variables to reinforce exercise recommendations in a low physical performance and overweight majors.

1. Introduction

The university or college lifestyle may generate changes in students’ health due to sedentary habits [1]. Some situations that lead to health problems among university students include inadequate nutrition, leisure time, low levels of physical activity, the consumption of psychoactive substances, stress mismanagement, academic and psychosocial pressure, and an insufficient follow-up of medical recommendations [2,3,4]. A high prevalence of sedentary behavior in university students of different majors has been reported which affects the diverse aspects of wellbeing and might result even in depression [5] or being overweight [6]. This in turn modifies the biological and psychological parameters that disrupt the quality of life. Besides the aforementioned factors, the initiation of university life leads to an increase in academic activities, social interaction with peers, and the constant use of electronic devices or screen time; therefore, a significantly reduced physical activity, increased intake of “fast food”, and raised stress levels might predispose this population to the development of chronic diseases [7].
In Colombia, evidence related to sedentary behavior among university students has shown that students spend several hours of their time on electronic devices per day [8]. A systematic review reported that sedentary behavior, especially the sitting time and screen exposure, is positively associated with obesity and other chronic diseases (e.g., dyslipidemia, hypertension, diabetes, etc.) in young Latin American adults [9]. Therefore, wellness programs should be considered to be an elementary component of tertiary education institutions to deepen the aspects of university students’ lifestyles, which directly impacts the lack of activity and increased sedentary behavior [10,11,12,13]. Additionally, it is important to highlight the academic load that students undergo each semester which influence the individual conditions and the leisure spaces. Actually, major students may consider physical activity as irrelevant or unrelated to their health and academic performance, which is contrary to the available evidence [14].
In undergraduate programs related to physical activity or sport (e.g., Physical Education or Sports Sciences majors), there are admission protocols that involve standardized physical tests that, in part, allow for evaluating the level of physical fitness (PF) of the applicants. In this sense, it has been found that PF is adequate in the first year. Still, it substantially deteriorates compared to senior students, which could be attributed to academic dynamics and workload, among other personal factors to which most future graduates are exposed to [2]. Despite the relevance of the PF evaluation process for the detection and control of the risk factors in the university population, monitoring this aspect is not a process that is carried out systematically in university training programs, including those of physical education, sports, or similar majors. Thus, this study aimed to identify the physical fitness profiles of majors enrolled on the Physical Culture program at the Universidad Santo Tomás in Bogotá, Colombia. We believe this reproducible analytical approach would provide valuable information for implementing assertive strategies for health promotion and disease prevention among university students.

2. Materials and Methods

2.1. Study Design

This cross-sectional study was reported following the Strengthening the Reporting of Observational Studies in Epidemiology—STROBE guidelines [15].

2.2. Setting

The study was carried out from February to November 2019. The standardization considered the following recommendations given to the participants to avoid biases in the data collection. Concerning their diet, the participants were instructed that they should not have ingested food two-four hours prior to the evaluation, nor should they have consumed alcohol or coffee eight hours before measuring their body mass. Similarly, they were informed that they should not have consumed corticosteroid or diuretic medications eight days before the assessment. Regarding physical activity/physical exercise, it was insisted that no physical exertion should be performed 24 h before attending the test; in addition, bladder emptying was suggested prior to the data collection. The physical tests were carried out in the San Alberto Magno Campus of the Santo Tomás University (Bogota, Colombia) under the supervision of researchers in physical activity with previous experience in applying these tests. The subjects voluntarily participated and were informed about this research’s protocol and aim. The study was designed following the ethical guidelines of the Declaration of Helsinki [16] and the Resolution 8439 of 1993 of the Ministry of Health of Colombia. The Ethics Committee approved the research protocol of the Universidad Manuela Beltrán (CEI-1705228-27).

2.3. Participants

The healthy and physically active undergraduate students of a Physical Education, Sports and Recreation major at a private institution in Bogota, Colombia, were invited to participate. This study belonged to a research project approved and developed in collaboration with other institutions. The selection of the subjects was based on the following inclusion criteria: (i) being enrolled in a Physical Education, Sports and Recreation program; (ii) residing in Bogotá and municipalities of the metropolitan area (above 2600 m above sea level); and (iii) not having any medical history or restriction that would make participation impossible during the execution of the physical tests. All the participants in this study were active undergraduate students of the Physical Culture, Recreation and Sport course at the Universidad Santo Tomás.

2.4. Variables

The following continuous variables were measured: the body mass (kg), stature (cm), cardiorespiratory fitness (indirect VO2max), muscle power (horizontal jump), sprint speed (30 m), agility (Shuttle Run Agility Test), and flexibility (Sit and Reach). Other variables were also collected from the informed consent (sex and academic year). The socioeconomic stratum was obtained from the Bogotá’s planning department using the official address of the participants’ residence.

2.5. Data Sources/Measurement

Bogotá city is 2630 m above sea level with temperatures ranging from 9 to 14 °C. The familiarization and data collection were performed in two sessions. In the first session, recommendations were given for the correct execution of the battery test. In addition, it was explained to the participants that one of the examiners would supervise the movements to avoid compromising technical execution while a second examiner would proceed with a photography collection (for comparison with the standardized images of the test). During the second session, the data collection was made between 07:00 and 10:00 and 18:00 and 21:00 (GMT-5), the times the participants were available. Prior to testing, the participants performed a general warm-up consisting of low-intensity jogging for 10 min. All tests were performed once, except for the horizontal jump, which had two attempts. The following order was maintained during the testing process.

2.5.1. Anthropometry

The body mass data were obtained using the Tanita® SC 331S (Tanita Corporation of America, Inc., Arlington Heights, IL, USA). The participants’ stature was measured to the nearest 0.1 cm with a portable stadiometer SECA® 213 (Medical Scales and Measuring Systems, Hamburg, Germany).

2.5.2. Cardiorespiratory Fitness

The maximum oxygen uptake (VO2max), as an indicator of cardiorespiratory fitness, was estimated using field tests: the Cooper Run Test (CRT) and Shuttle Run Test (SRT-20m). The CRT protocol was performed on an athletics track (400 m), marked every 100 m, to obtain the distance each subject covered during the 12 min of the test [17]. For this test, the participants were scheduled in the morning (between 8 and 10 A.M., GMT-5), when the sunlight exposure would not be so intense as to reduce the physical performance and thus biasing the results. For the estimation of VO2max (in mL·kg−1·min−1), we used the equation 22.351 × distance (km) − 11.288. This was developed by Cooper [17] and validated by Penry et al. [18]. On the other hand, the SRT-20m was carried out on the same schedule, but eight days after the application of the CRT to give the participants time to recover. This test was performed in an open field with a flat and stable surface where two cones were installed with 20 m between them. The participants stood next to one of these cones and covered the distance mentioned above in a round trip. The initial velocity was 8.5 km·h−1, and the pace of the race increased by 0.5 km·h−1 every minute, indicated by a sound previously explained to the individuals. The test ended once the participant failed to keep up with the speed (i.e., 3 m behind the 20 m line on the audio cue) or was able to complete the stage. With the results of this test, the equation 5.857 × final velocity (km/h) − 19.458 [19] was used for the estimation of the VO2max (in mL·kg−1·min−1).

2.5.3. Lower-Limb Power

The horizontal jump test without a pre-run was performed on a running track with a 300 cm tape measure placed on the ground; the distance reached from 0 cm to the heel of the foot closest to 0 cm was recorded. The protocol was based on the study of Manouras et al. [20] to evaluate the lower-limb muscle power of the participants. Some attempts were made to become familiar with the test, as established by the committee of experts who created the EUROFIT battery [21]. It is worth noting that the validation of different jumping tests using Alpha Cronbach’s coefficients by Markovic et al. [22] showed high-reliability values for the horizontal jump tests (0.93–0.96). A rest interval of two to three minutes was afforded between the three attempts. The highest recorded value was used for the analysis.

2.5.4. Sprint Speed

The 30 m test in the static position was chosen, which was applied on an athletic track with a flat and consistent surface; cones delimited the distance, and the time data obtained during the execution was recorded (Polar RS100; Polar Electro Oy, Kempele, Finland). The purpose of this test is to evaluate the reaction speed and acceleration of the participant. It is a differentiated test compared to other tests that measure the displacement speed. A level of reliability between 0.88 and 0.95 has been found, which varies according to the age and the terrain used, preferably on an athletic track [23].

2.5.5. Agility

The Shuttle Run Agility Test (SRAT) was applied in an open field, with a delimitation of cones at 9.14 m from the initial line, where the wooden blocks were transported in the respective order of the test. Three changes in direction of 180° were made; the action of transporting and covering this length should be in the shortest possible time (Polar RS100; Polar Electro Oy, Kempele, Finland), according to the physical condition of each subject [24]. This agility test was compared and validated by Kutlu et al. [25] who reported no differences between the SRAT versus other agility tests such as the Illinois protocol, zig-zag test, 30 m, Bosco, and T-drill Agility.

2.5.6. Flexibility

The participants’ flexibility was assessed using the classic Sit and Reach (SR) test. In general, SR tests are valid for estimating the hamstring’s extensibility, although they have a low mean validity for estimating the lumbar’s extensibility [26]. They have a high relative intra-rater reliability (intra-class correlation coefficient of 0.89–0.99) [27].

2.6. Study Size

Non-probability convenience sampling was used. After the announcement to participate in this study, only the college students that fulfilled all the inclusion criteria were considered enrolled. A total of 613 students were suitable for eligibility.

2.7. Statistical Analysis

Descriptive statistics were expressed as the mean, standard deviation, and 95% confidence intervals (95% CI). The data distribution was plotted using horizontal half-violin diagrams. The data were analyzed with the Mann–Whitney U and Kruskal–Wallis (Bonferroni post hoc) tests to determine the differences between the sexes and between the academic years and socioeconomic strata, respectively. Eta-squared (η2) was used to report the magnitude of differences assuming 0.09, 0.14, and >0.22 as a small, medium, and large effect size [28]. As we have performed previously [29,30,31,32], the participants were subdivided into clusters using unsupervised machine learning to identify similar data points (natural groupings) and extract the profile patterns. We used the partitioning around the medoids (PAM) algorithm, also known as k-Medoids clustering which, unlike the k-means algorithm, considers the median as the center of a cluster, thus, it is more robust to noises and outliers [33]. This has enhanced the robustness against the outliers and reduced the noise in the unsupervised machine learning process [33]. Moreover, we performed hierarchical clustering with the bottom-up approach, which provides an easy-to-interpret view of the clustering structure [34]. The number of clusters was determined using 30 criterion algorithms comparing the two-to-ten cluster solutions with the R package ‘NbClust’. The internal validation for selecting the clustering method to discuss our results was performed with the ‘clValid’ package [35]. To determine the differences between the clusters, we performed a Mann–Whitney U test. The packages ‘factoextra’ and ‘fmsb’ were used to visualize the clustering results and represent the comparison between the clusters as a spider plot, respectively, within the free software environment for the statistical computing and graphics R v4.0.2 [36]. A significance level of p < 0.05 was considered using the IBM SPSS v26 (IBM Corp., Armonk, NY, USA).

3. Results

3.1. Participants

Seventy-one students did not complete all the battery of the tests and were excluded. Therefore, we report the analysis with data obtained for 542 majors attending from the 1st to 4th year of the physical education major (445 males and 97 females). Differences were found between the participants’ age, body mass, and height by sex, but not in the body mass index (Table 1).
The distribution of the physical fitness variables by sex is represented in Figure 1. According to the results, differences were found in all the study variables by sex (p < 0.05). Males showed higher levels of cardiorespiratory fitness, lower limb power, and speed; however, university females showed higher levels of flexibility (Figure 1).
The characteristics of the population by academic year and by socioeconomic strata are shown in Table 2.
As expected, statistically significant age differences were found between the freshmen compared to other students (sophomore, junior, and senior). Significantly lower CRT distance covered and CRT VO2max values were seen in seniors against freshman and sophomore students. In the case of the SRT-20M final stage velocity and SRT-20M VO2max, there were differences between freshman and junior, sophomore and junior, and junior and sophomore. The 30 m sprint showed lower values in sophomores and seniors compared to freshmen. Although statistically non-significant, the SRAT revealed that juniors (third year students) were less agile than other students. When data are grouped by socioeconomic strata, a higher age was observed in the students classified in medium versus low strata (Table 2).

3.2. Main Results

The clValid() function revealed that the hierarchical clustering analysis was more appropriate to cluster our data than other methods. Two clusters were identified: n = 318 (cluster 1) and n = 224 (cluster 2), as can be seen in Figure 2.
Physical fitness accounts for the variation in the data. In fact, the comparison between the clusters revealed significant differences in all the variables except for the stature and flexibility (Table 3). This allowed for recognizing clustering as a relevant methodology to describe the differences between the identified phenotypes (clusters) and help to better ensure an internal validity. Multidimensional analysis showed that the body mass, agility, and CRT VO2max were the variables that explained most of the variance in the data (Figure 3).

4. Discussion

4.1. Key Results

This study aimed to profile the physical fitness of undergraduate students of a physical education major using unsupervised machine learning. We were able to identify two physical fitness-based profiles of majors: (i) older, heavier, overweight, less VO2max, less lower-limb power, faster, less agile, and less flexible students (Cluster 1); and (ii) younger, taller, normal weight, higher VO2max, more lower-limb power, slower, more agile, and more flexible students (Cluster 2). The main results also showed that as the major progress, the cardiorespiratory fitness decreases, especially in the final year (senior) students. Castro et al. (2020) [2] also observed that agility, speed, and flexibility decrease as the academic curricula advances, with an accentuation in the last year. A decreased cardiorespiratory fitness is associated with sedentary habits in college students [37,38]. It is worth noting that speed showed better outcomes in sophomores and seniors compared to freshmen (first-year students). We have previously reported the reference values for indirect VO2max tests (CRT and SRT-20m) at a high altitude—between 2600 and 3700 m above sea level—because we are aware of the incidence of this environmental condition on the oxygen supply, oxygen utilization, air density, as well as the cellular response [39]. The participants of this study are within this context since a high altitude is considered to be those locations between 2000 to 5000 m above the sea level [40]. Our results also revealed that female participants had a lower cardiorespiratory fitness when compared to males (Figure 1). This significant difference could be partially explained by differences in the serum testosterone concentration if we consider the documented erythrogenic effects of this hormone [41].
Interestingly, the body mass, agility, and cardiovascular fitness explained most of the data variance among the generated profiles (clusters). This might account for the phenotypic differences. We recently reported that applicants to a physical education major have a normal body composition and body mass index values, which indicates that the nutritional status in the first academic years is appropriate for the specific energy needs of the major (which includes heavy physical training) [32]. Notwithstanding, further research should evaluate if the decreased physical condition that we found in senior students may be associated with alterations in the body composition and health-related parameters. Nowadays, it is clear that an unhealthy lifestyle (e.g., overweight/obese phenotype) can trigger long-term health problems and chronic diseases [42,43,44]. Moreover, Lee and Kim [5] highlighted that a sedentary lifestyle increases anxiety, stress, and depression in university and college students.
Complementary, there is a significant association between a low physical fitness and a poor academic performance [45]; therefore, it becomes important to promote physical and recreational activities during university life. In fact, pre-established university activities, facilities, and the accurate identification of students at a higher risk are important factors that might positively impact the physical fitness level, quality of life, and academic performance of major students [46]. Here, we presented a methodological and analytical workflow to profile exercise and sports major students using unsupervised machine learning. This data analysis approach goes beyond the classical descriptive statistical analysis and is based on various algorithms that can be chosen depending on the nature of the data [47]. Our research group used this methodology to generate the profiles of young athletes and college students based on morphological and functional data [29,30,31]. These analytical methodologies are highly reproducible by the scientific community and help to extract further information from a given dataset that might be used to explain the contextualized phenomena in different ways [29]. Concerning the application of this data analysis approach in universities, we encourage human resources and health departments to become familiarized with and apply more frequently unsupervised machine learning algorithms to identify profiles in college students. An integral framework that can be used for assessing wellbeing in majors is the PERMA+4 [48]. This framework is based on nine building blocks to develop wellbeing: positive emotions, engagement, relationships, meaning, accomplishment, physical health, mindset, environment, and economic security. For more detailed coverage of this model, please refer to Donaldson et al. [49].
Our study has certain limitations that should be considered when drawing practical inferences and setting up future research. First, the small sample size of the female population belonging to the physical education major makes it difficult to analyze this subgroup by ranges of ages or academic semesters. Second, since we did not record the body composition, athletic experience, or academic success variables, more research is needed to find associations with physical fitness in this type of student. Third, observational studies cannot be used to demonstrate causality; therefore, future experimental studies are needed to evaluate the effect of modifying variables that explain most of the variation in our clusters.

4.2. Interpretation and Generalizability

This observational study provides information regarding the physical fitness of physical education majors residing in a high altitude (above 2600 m above sea level). It also reports potential variables that account for data variation between the majors’ profiles (body mass, agility, and cardiorespiratory fitness). Finally, this study also contributes to the analytical procedures to profile data under the machine learning paradigm and establish rationale for future research. In fact, applying machine learning to characterize undergraduate populations contributes to the identification of patterns, as well as providing a reproducible, feedback-based, and fast tool to analyze data.

5. Conclusions

A decreased physical fitness was found in undergraduate students enrolled in a physical education major as the academic program progressed. This is consistent with the literature, but interestingly, only their speed improved as their college life advanced. In addition, there were no significant differences in the cardiorespiratory fitness, power, agility, or flexibility based on socioeconomic stratification. We identified two significantly different phenotypes representing majors with a poor and good physical performance. The information contained in this article might help universities in the early identification of at-risk students and to develop educational programs to promote healthy lifestyles during college age.

Author Contributions

Conceptualization, D.A.B. and I.A.S.-R.; data collection, I.A.S.-R. and D.M.-R.; formal analysis, D.A.B.; software and visualization, D.A.B.; writing—original draft, D.A.B., I.A.S.-R. and J.L.P.; writing—review and editing, Y.M., J.K., L.M.G.-M., D.R.-V. and R.B.K. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the ‘Facultad de Cultura Física y Deporte’ (grant number CEI-1705228-27) at Universidad Santo Tomás (Bogotá, Colombia). The APC was funded by the Exercise and Sport Nutrition Laboratory (Texas A&M).

Institutional Review Board Statement

The study was conducted in accordance with the declaration of Helsinki and the protocol was approved by the Ethics Committee of the Universidad Manuela Beltrán.

Informed Consent Statement

Written informed consent was obtained from all participants.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon request.

Acknowledgments

The authors would like to thank the undergraduate students for their support and participation in this research project. We also thank Angie Rivera Joven and Luigi Roa Peralta for their assistance during the data collection.

Conflicts of Interest

D.A.B. has conducted academic-sponsored research on resistance training and has received honoraria for selling linear position transducers and speaking about exercise sciences at international conferences and private courses. R.B.K. has conducted industry-sponsored studies at the universities he has been affiliated with and occasionally serves as a scientific and legal consultant related to exercise and nutrition intervention studies. The other authors declare no conflicts of interest. All authors are responsible for the content of this article.

References

  1. Arévalo, M.T.V.; Muñoz, A.F.O.; Cuevas, J.R.T. Tipologías de estilos de vida en jóvenes universitarios. Univ. y Salud 2016, 18, 246–256. [Google Scholar] [CrossRef] [Green Version]
  2. Castro Jiménez, L.E.; Vásquez Cubillos, N.; Tovar Yate, H.A.; Valderrama Sánchez, J.A.; Argüello Gutiérrez, Y.P. Cambios en la condición física de jóvenes aparentemente sanos desde su ingreso a la universidad. VIREF Rev. De Educ. Física 2020, 8, 91–102. [Google Scholar]
  3. Dodd, L.J.; Al-Nakeeb, Y.; Nevill, A.; Forshaw, M.J. Lifestyle risk factors of students: A cluster analytical approach. Prev. Med. 2010, 51, 73–77. [Google Scholar] [CrossRef] [PubMed]
  4. Deliens, T.; Deforche, B.; De Bourdeaudhuij, I.; Clarys, P. Determinants of physical activity and sedentary behaviour in university students: A qualitative study using focus group discussions. BMC Public Health 2015, 15, 201. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  5. Lee, E.; Kim, Y. Effect of university students’ sedentary behavior on stress, anxiety, and depression. Perspect. Psychiatr. Care 2019, 55, 164–169. [Google Scholar] [CrossRef] [Green Version]
  6. Chaves-Franco, D.; Lujan-Ferraz, N.; Ferreira de Sousa, T. Sedentary behavior among university students: A systematic review. Rev. Bras. De Cineantropometria 2019, 21, e56485. [Google Scholar] [CrossRef]
  7. Al-Nakeeb, Y.; Lyons, M.; Dodd, L.J.; Al-Nuaim, A. An Investigation into the Lifestyle, Health Habits and Risk Factors of Young Adults. Int. J. Environ. Res. Public Health 2015, 12, 4380–4394. [Google Scholar] [CrossRef] [Green Version]
  8. Sánchez-Guette, L.; Herazo-Beltrán, Y.; Galeano-Muñoz, L.; Romero-Leiva, K.; Guerrero-Correa, F.; Mancilla-González, G.; Pacheco-Rodríguez, N.; Ruiz-Marín, A.; Pino, L.O. Comportamiento sedentario en estudiantes universitario. Rev. Latinoam. De Hipertens. 2019, 14, 393–397. [Google Scholar]
  9. Natali, A.J.; Marin, K.A.; Hermsdorf, H.H.M.; Rezende, F.A.C.; Peluzio, M.D.C.G. A systematic review of cross-sectional studies on the association of sedentary behavior with cardiometabolic diseases and related biomarkers in South American adults. Nutr. Hosp. 2020, 37, 359–373. [Google Scholar] [CrossRef] [Green Version]
  10. Kellner, M.; Faas, F. Get up, stand up: A randomized controlled trial to assess the effectiveness of a messenger-based intervention to reduce sedentary behavior in university students. Z Gesundh Wiss 2022, 1–9. [Google Scholar] [CrossRef]
  11. Perez Lopez, I.J.; Rivera Garcia, E.; Delgado-Fernandez, M. Improvement of healthy lifestyle habits in university students through a gamification approach. Nutr. Hosp. 2017, 34, 942–951. [Google Scholar] [CrossRef] [PubMed]
  12. Paredes, A.F. Efectos del programa de actividad física y deportes en estudiantes de medicina. Comuni@cción Rev. De Investig. En Comun. Y Desarro. 2020, 11, 142–152. [Google Scholar] [CrossRef]
  13. Worobetz, A.; O’Callaghan, M.; Walsh, J.; Casey, M.; Hayes, P.; Bengoechea, E.G.; Woods, C.; McGrath, D.; Glynn, L.G. Exercise Compared to Mindfulness for Physical and Mental Wellbeing in Medical Students. Ir. Med. J. 2022, 115, 560. [Google Scholar]
  14. Chen, H.; Zhang, G.; Wang, Z.; Feng, S.; Li, H. The Associations between Daytime Physical Activity, While-in-Bed Smartphone Use, Sleep Delay, and Sleep Quality: A 24-h Investigation among Chinese College Students. Int. J. Environ. Res. Public Health 2022, 19, 9693. [Google Scholar] [CrossRef]
  15. Vandenbroucke, J.P.; Von Elm, E.; Altman, D.G.; Gøtzsche, P.C.; Mulrow, C.D.; Pocock, S.J.; Poole, C.; Schlesselman, J.J.; Egger, M. Mejorar la comunicación de estudios observacionales en epidemiología (STROBE): Explicación y elaboración. Gac. Sanit. 2009, 23, 158.e1–158.e28. [Google Scholar] [CrossRef]
  16. World Medical Association. World Medical Association Declaration of Helsinki. Ethical principles for medical research involving human subjects. Nurs. Ethics 2002, 9, 105–109. [Google Scholar] [CrossRef]
  17. Cooper, K.H. A means of assessing maximal oxygen intake. Correlation between field and treadmill testing. JAMA 1968, 203, 201–204. [Google Scholar] [CrossRef]
  18. Penry, J.T.; Wilcox, A.R.; Yun, J. Validity and Reliability Analysis of Cooper’s 12-Minute Run and the Multistage Shuttle Run in Healthy Adults. J. Strength Cond. Res. 2011, 25, 597–605. [Google Scholar] [CrossRef] [PubMed]
  19. Léger, L.A.; Lambert, J. A maximal multistage 20-m shuttle run test to predict VO2 max. Eur. J. Appl. Physiol. Occup. Physiol. 1982, 49, 1–12. [Google Scholar] [CrossRef] [PubMed]
  20. Manouras, N.; Papanikolaou, Z.; Karatrantou, K.; Kouvarakis, P.; Gerodimos, V. The efficacy of vertical vs. horizontal plyometric training on speed, jumping performance and agility in soccer players. Int. J. Sports Sci. Coach. 2016, 11, 702–709. [Google Scholar] [CrossRef]
  21. Council of Europe. Handbook for the Eurofit Test on Physical Fitness; Council of Europe.: Strasbourg, France, 1993. [Google Scholar]
  22. Markovic, G.; Dizdar, D.; Jukic, I.; Cardinale, M. Reliability and factorial validity of squat and countermovement jump tests. J. Strength Cond. Res. 2004, 18, 551–555. [Google Scholar] [CrossRef] [PubMed]
  23. Altmann, S.; Ringhof, S.; Neumann, R.; Woll, A.; Rumpf, M.C. Validity and reliability of speed tests used in soccer: A systematic review. PLoS ONE 2019, 14, e0220982. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  24. Tubagi-Polito, L.F.; de Moura Carneiro, Y.; Aparecida-Moscaleski, L.; Figueira-Junior, A.J.; Callegari-Zanetti, M.; Giuliano de Sá Pinto, C.; Magalhães-Dias, H.; Rodrigues-Nunes, H.; Inácio de Lima, S. Shuttle Run Agility Test in Soccer Athletes of Under–10 Category with Dry and Wet Conditions Field. Int. J. Sport. Sci. Coach. 2017, 7, 45–49. [Google Scholar]
  25. Kutlu, M.; Yapıcı, H.; Yoncalık, O.; Çelik, S. Comparison of a New Test For Agility and Skill in Soccer With Other Agility Tests. J. Hum. Kinet. 2012, 33, 143–150. [Google Scholar] [CrossRef]
  26. Mayorga-Vega, D.; Merino-Marban, R.; Viciana, J. Criterion-Related Validity of Sit-and-Reach Tests for Estimating Hamstring and Lumbar Extensibility: A Meta-Analysis. J. Sports Sci. Med. 2014, 13, 1–14. [Google Scholar] [PubMed]
  27. Ayala, F.; de Baranda, P.S.; Croix, M.D.S.; Santonja, F. Fiabilidad y validez de las pruebas sit-and-reach: Revisión sistemática. Rev. Andal. de Med. del Deport. 2012, 5, 57–66. [Google Scholar] [CrossRef] [Green Version]
  28. Salkind, N.J. Eta-Squared. In Encyclopedia of Research Design; SAGE Publications, Inc.: Thousand Oaks, CA, USA, 2010; Volume 1. [Google Scholar]
  29. Bonilla, D.A.; Peralta, J.O.; Bonilla, J.A.; Urrutia-Mosquera, W.; Vargas-Molina, S.; Cannataro, R.; Petro, J.L. Morphology, body composition and maturity status of young Colombian athletes from the Urabá subregion: A k-Medoids and hierarchical clustering analysis. J. Hum. Sport Exerc. 2020, 15, 20. [Google Scholar] [CrossRef]
  30. Bonilla, D.A.; Franco-Hoyos, K.; Agudelo-Martínez, A.; Kammerer-López, M.; Bedoya-Bedoya, L.F.; Moreno, Y.; Petro, J.L. Clustering-based characterization of body composition and morphology in Colombian nutrition and dietetics undergraduate students (CES University, 2016-2020). In Proceedings of the 4th International Symposium of Advanced Topics in Exercise Physiology: Non-Pharmacological Treatment for the Improvement of the Quality of Life in the Elderly, Ensenada, México, 3–5 November 2020. [Google Scholar]
  31. Bonilla, D.A.; Peralta-Alzate, J.O.; Bonilla-Henao, J.A.; Urrutia-Mosquera, W.; Cannataro, R.; Kočí, J.; Petro, J.L. Unsupervised machine learning analysis of the anthropometric characteristics and maturity status of young Colombian athletes. J. Phys. Educ. Sport 2022, 22, 256–265. [Google Scholar] [CrossRef]
  32. Gasperín-Rodríguez, E.I.; Gómez-Figueroa, J.A.; Gómez-Miranda, L.M.; Ríos-Gallardo, P.T.; Palmeros-Exsome, C.; Hernández-Lepe, M.A.; Moncada-Jiménez, J.; Bonilla, D.A. Body Composition Profiles of Applicants to a Physical Education and Sports Major in Southeastern Mexico. Int. J. Environ. Res. Public Health 2022, 19, 15685. [Google Scholar] [CrossRef]
  33. Mannor, S.; Jin, X.; Han, J.; Jin, X.; Han, J.; Jin, X.; Han, J.; Zhang, X. K-Medoids Clustering. In Encyclopedia of Machine Learning; Spinger: Berlin/Heidelberg, Germany, 2011; pp. 564–565. [Google Scholar]
  34. Wittek, P. Unsupervised Learning. In Quantum Machine Learning; Elsevier: Amsterdam, The Netherlands, 2014; pp. 57–62. [Google Scholar]
  35. Brock, G.; Pihur, V.; Datta, S.; Datta, S. clValid: AnRPackage for Cluster Validation. J. Stat. Softw. 2008, 25, 1–22. [Google Scholar] [CrossRef] [Green Version]
  36. R Core Team. R: A Language and Environment for Statistical Computing, Version 4.0.3; R Foundation for Statistical Computing: Vienna, Austria, 2020. [Google Scholar]
  37. Lepp, A.; Barkley, J.E.; Sanders, G.J.; Rebold, M.; Gates, P. The relationship between cell phone use, physical and sedentary activity, and cardiorespiratory fitness in a sample of U.S. college students. Int. J. Behav. Nutr. Phys. Act. 2013, 10, 79. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  38. Calestine, J.; Bopp, M.; Bopp, C.M.; Papalia, Z. College Student Work Habits are Related to Physical Activity and Fitness. Int. J. Exerc. Sci. 2017, 10, 1009–1017. [Google Scholar] [PubMed]
  39. Sánchez-Rojas, I.A.; Ejercicio, D.Y.S.-.G.G.d.I.C.A.a.; Romero, D.M.; Argüello-Gutiérrez, Y.P.; Castro-Jiménez, L.E.; Triana-Reina, H.R.; Perez-Cebreros, E.A.; Petro, J.L.; Bonilla, D.A. Colombia Valores de referencia para las pruebas de Cooper y de 20m de ida y vuelta en población residente en altitud elevada. [Reference values for Cooper and Shuttle Run Tests in population living at high altitude]. RICYDE. Rev. Int. de Cienc. del Deport. 2021, 17, 221–233. [Google Scholar] [CrossRef]
  40. Sinex, J.A.; Chapman, R.F. Hypoxic training methods for improving endurance exercise performance. J. Sport Health Sci. 2015, 4, 325–332. [Google Scholar] [CrossRef]
  41. Gabrielsen, J.S. Iron and Testosterone: Interplay and Clinical Implications. Curr. Sex. Health Rep. 2017, 9, 5–11. [Google Scholar] [CrossRef]
  42. González, K.; Fuentes, J.; Márquez, J.L. Physical Inactivity, Sedentary Behavior and Chronic Diseases. Korean J. Fam. Med. 2017, 38, 111–115. [Google Scholar] [CrossRef] [Green Version]
  43. Anderson, E.; Durstine, J.L. Physical activity, exercise, and chronic diseases: A brief review. Sports Med. Health Sci. 2019, 1, 3–10. [Google Scholar] [CrossRef]
  44. Laaksonen, D.E.; Lakka, H.-M.; Salonen, J.T.; Niskanen, L.K.; Rauramaa, R.; Lakka, T.A. Low Levels of Leisure-Time Physical Activity and Cardiorespiratory Fitness Predict Development of the Metabolic Syndrome. Diabetes Care 2002, 25, 1612–1618. [Google Scholar] [CrossRef] [Green Version]
  45. de Greeff, J.W.; Bosker, R.J.; Oosterlaan, J.; Visscher, C.; Hartman, E. Effects of physical activity on executive functions, attention and academic performance in preadolescent children: A meta-analysis. J. Sci. Med. Sport 2018, 21, 501–507. [Google Scholar] [CrossRef]
  46. Pedersen, M.R.L.; Bredahl, T.V.G.; Elmose-Østerlund, K.; Hansen, A.F. Motives and Barriers Related to Physical Activity within Different Types of Built Environments: Implications for Health Promotion. Int. J. Environ. Res. Public Health 2022, 19, 9000. [Google Scholar] [CrossRef]
  47. Von Ziegler, L.; Sturman, O.; Bohacek, J. Big behavior: Challenges and opportunities in a new era of deep behavior profiling. Neuropsychopharmacology 2020, 46, 33–44. [Google Scholar] [CrossRef] [PubMed]
  48. Kočí, J. Nový Pohled na Zdraví a Psychický Well-Being—PERMA. In Zdraví a Mentální Well-Being Student; CRP Rozcestník: New York, NY, USA, 2021. [Google Scholar]
  49. Donaldson, S.I.; van Zyl, L.E.; Donaldson, S.I. PERMA+4: A Framework for Work-Related Wellbeing, Performance and Positive Organizational Psychology 2.0. Front. Psychol. 2021, 12, 817244. [Google Scholar] [CrossRef] [PubMed]
Ijerph 20 00146 g001 550
Figure 1. Horizontal half violin plots of the physical fitness variables by sex. This figure shows the distribution of the different physical fitness variables measured in female (pink) and male (blue) majors. Black circles represent the mean, while dark grey diamonds the raw data. CRT: Cooper Run Test; SRAT: Shuttle Run Agility Test; SRT-20M: Shuttle Run Test. * Significant difference p < 0.05.
Figure 1. Horizontal half violin plots of the physical fitness variables by sex. This figure shows the distribution of the different physical fitness variables measured in female (pink) and male (blue) majors. Black circles represent the mean, while dark grey diamonds the raw data. CRT: Cooper Run Test; SRAT: Shuttle Run Agility Test; SRT-20M: Shuttle Run Test. * Significant difference p < 0.05.
Ijerph 20 00146 g001
Ijerph 20 00146 g002a 550Ijerph 20 00146 g002b 550
Figure 2. Clustering plots of the analyzed data. (A). Dendrogram of the bottom-up agglomerative clustering. Each leaf corresponds to one student. Students who are similar are combined into branches, which are fused at a higher height. The height of the fusion, provided on the vertical axis, indicates the (dis)similarity/distance between two students/clusters. The hierarchical tree was cut to partition the data into clusters (blue for Cluster 1 and red for Cluster 2). (B). Cluster plot of the k-Medoids analysis. Red for Cluster 1 and blue for Cluster 2.
Figure 2. Clustering plots of the analyzed data. (A). Dendrogram of the bottom-up agglomerative clustering. Each leaf corresponds to one student. Students who are similar are combined into branches, which are fused at a higher height. The height of the fusion, provided on the vertical axis, indicates the (dis)similarity/distance between two students/clusters. The hierarchical tree was cut to partition the data into clusters (blue for Cluster 1 and red for Cluster 2). (B). Cluster plot of the k-Medoids analysis. Red for Cluster 1 and blue for Cluster 2.
Ijerph 20 00146 g002aIjerph 20 00146 g002b
Ijerph 20 00146 g003 550
Figure 3. Multidimensional 3D plot of pre-defined variables. The arrangement of clusters within a 3D structure enables to identify of the significant variables that contribute the most to data variation by the natural generation of subspaces.
Figure 3. Multidimensional 3D plot of pre-defined variables. The arrangement of clusters within a 3D structure enables to identify of the significant variables that contribute the most to data variation by the natural generation of subspaces.
Ijerph 20 00146 g003
Table
Table 1. Characteristics of the participants.
Table 1. Characteristics of the participants.
VariableMale (n = 445)95% CIFemale (n = 97)95% CIp-Valueη2
Age20.0 (2.2)19.7–20.219.4 (2.2)18.9–19.8<0.0010.013
BM (kg)67.8 (9.8)66.9–68.7757.8 (8.6)56.0–59.5<0.0010.144
Stature (cm)171.7 (6.2)171.2–172.3159.2 (5.9)158.0–160.4<0.0010.324
BMI (kg·m−2)22.9 (2.7)22.6–23.222.7 (2.9)22.1–23.30.2180.002
Data are expressed as mean (SD). BM: body mass; BMI: body mass index. The statistically significant differences at a level of 0.05 for the Mann–Whitney U test are shown. Effect size as eta-squared (η2).
Table
Table 2. Characteristics of the participants grouped by academic year and socioeconomic strata.
Table 2. Characteristics of the participants grouped by academic year and socioeconomic strata.
VariableFreshmanSophomoreJuniorSenior
(n = 279)(n = 132)(n = 59)(n = 72)
Age 18.8 (1.8) a,b,c20.5 (2.1) e20.2 (1.3) f22.3 (2.2)
Body mass (kg)65.9 (10.4)66.0 (10.2)65.7 (9.8)66.8 (10.4)
Stature (cm)169.8 (8.1)169.0 (7.0)170.9 (8.7)169.8 (7.2)
BMI (kg·m−2)22.9 (2.7)23.0 (3.0)22.4 (2.2)23.1 (2.7)
CRT Distance covered (km)2.30 (0.4) c2.33 (0.4) e2.20 (0.4)2.09 (0.4)
CRT VO2max40.1 (9.5) c40.8 (10.5) e38.0 (10.1)35.7 (10.0)
SRT-20M Final stage velocity (km·h−1)11.8 (1.1) b12.0 (0.8) d12.8 (1.4) f11.6 (0.9)
SRT-20M VO2max49.7 (6.5) b50.8 (5.0) d55.8 (8.4) f48.9 (5.7)
Horizontal
jump (cm)		195.4 (30.1)202.2 (34.6)198.7 (21.7)193.2 (32.3)
Sprint 30 m (s)8.0 (2.0) a,c6.5 (2.2) d7.8 (0.7) f6.6 (2.0)
SRAT (s)9.1 (2.1) a,b10.1 (1.4) e10.2 (0.7) f9.2 (1.3)
Sit and reach test (cm)4.9 (8.3)5.3 (9.5)9.2 (8.1)5.2 (8.6)
VariableSS1
(n = 2)		SS2
(n = 89)		SS3
(n = 347)		SS4
(n = 90)		SS5
(n = 10)		SS6
(n = 4)
Age18.5 (2.1)19.2 (2.0) *19.9 (2.2)20.2 (2.3)19.8 (1.3)21.7 (2.9)
Body mass (kg)69.1 (1.5)65.0 (10.9)65.8 (9.9)68.0 (11.6)65.9 (6.7)64.4 (7.0)
Stature (cm)173.5 (2.1)169.2 (8.9)169.2 (7.3)171.0 (8.3)166.7 (6.1)169.5 (7.0)
BMI (kg·m−2)22.9 (0.0)22.6 (2.7)22.9 (2.7)23.1 (2.7)23.8 (3.0)22.4 (2.6)
CRT Distance covered (km)2.67 (0.1)2.32 (0.5)2.25 (0.4)2.27 (0.3)2.41 (0.4)2.39 (0.04)
CRT VO2max48.5 (2.3)40.6 (11.1)39.0 (10.2)39.5 (8.4)42.6 (9.5)42.1 (0.8)
SRT-20M Final stage velocity (km·h−1)12.7 (1.0)12.2 (1.3)11.9 (1.0)11.7 (1.0)11.5 (0.8)11.1 (0.4)
SRT-20M VO2max55.2 (6.2)52.0 (7.8)50.5 (6.4)49.5 (6.3)47.8 (5.1)45.7 (2.8)
Horizontal jump (cm)229.5 (17.6)195.9 (33.4)196.7 (30.4)197.9 (27.3)208.8 (40.1)203.2 (62.5)
Sprint 30 m (s)5.3 (2.3)7.6 (2.2)7.5 (2.0)7.3 (2.1)7.2 (2.6)7.6 (1.4)
SRAT (s)10.5 (0.9)9.5 (1.9)9.4 (1.7)9.4 (2.1)10.1 (1.3)10.3 (0.6)
Sit and reach test (cm)4.5 (3.5)5.6 (8.1)5.4 (8.7)5.3 (9.0)9.1 (10.1)14.7 (2.8)
Data are expressed as mean (SD). The social stratification in Colombia includes six strata, as follows: SS1 is lower-low, SS2 is low, SS3 is upper-low, SS4 is medium, SS5 is medium-high, and SS6 is high. Estimated VO2max expressed in mL·kg−1·min−1. CRT: Cooper Run Test; SRAT: Shuttle Run Agility Test; SRT-20m: Shuttle Run Test; VO2max: Maximum oxygen uptake. p-value < 0.05 for the Kruskal–Wallis test (Bonferroni post hoc). a Difference between freshman and sophomore; b difference between freshman and junior; c difference between freshman and senior; d difference between sophomore and junior; e difference between sophomore and senior; f difference between junior and senior; * difference between SS2 and SS4.
Table
Table 3. Characteristics of the profiled phenotypes.
Table 3. Characteristics of the profiled phenotypes.
VariableCluster 1
(n = 318)		95% CICluster 2
(n = 224)		95% CIp-Valueη2
Age (years)20.1 (2.5)19.8–20.419.5 (1.6)19.2–19.7<0.0010.008
Body mass (kg)67.6 (11.4)66.3–68.963.7 (8.0)63.5–64.8<0.0010.031
Stature (cm)169.1 (8.5)168.1–170.0170.1 (6.5)169.2–171.00.5590.000
BMI (kg∙m−2)23.2 (2.9)23.2–23.822.0 (2.2)21.7–22.2<0.0010.075
CRT Distance covered (km)2.10 (0.5)2.06–2.152.50 (0.3)2.45–2.55<0.0010.174
CRT VO2max35.8 (9.3)34.8–36.944.6 (8.7)43.4–45.7<0.0010.174
SRT Final stage velocity (km·h−1)11.3 (0.8)11.2–11.412.8 (0.9)12.6–12.9<0.0010.418
SRT VO2max47.0 (5.0)46.5–47.655.5 (5.4)54.8–56.2<0.0010.418
Horizontal jump (cm)194.6 (32.8)191.0–198.2200.7 (27.6)197.1–204.3<0.0010.008
Sprint (s)6.97 (2.1)6.73–7.218.21 (1.8)7.97–8.45<0.0010.085
SRAT (s)10.1 (1.4)9.97–10.298.65 (1.9)8.39–8.91<0.0010.162
Sit and reach test (cm)5.02 (8.8)4.04–6.006.34 (8.4)5.23–7.460.1150.004
Data are expressed as mean (SD) with the corresponding 95% confidence interval (95% CI). CRT: Cooper Run Test; SRAT: Shuttle Run Agility Test; SRT-20M: Shuttle Run Test; VO2max: maximum oxygen uptake. The statistically significant differences at a level of 0.05 for the Mann–Whitney U test are shown. Effect size as eta-squared (η2).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Bonilla, D.A.; Sánchez-Rojas, I.A.; Mendoza-Romero, D.; Moreno, Y.; Kočí, J.; Gómez-Miranda, L.M.; Rojas-Valverde, D.; Petro, J.L.; Kreider, R.B. Profiling Physical Fitness of Physical Education Majors Using Unsupervised Machine Learning. Int. J. Environ. Res. Public Health 2023, 20, 146. https://doi.org/10.3390/ijerph20010146

AMA Style

Bonilla DA, Sánchez-Rojas IA, Mendoza-Romero D, Moreno Y, Kočí J, Gómez-Miranda LM, Rojas-Valverde D, Petro JL, Kreider RB. Profiling Physical Fitness of Physical Education Majors Using Unsupervised Machine Learning. International Journal of Environmental Research and Public Health. 2023; 20(1):146. https://doi.org/10.3390/ijerph20010146

Chicago/Turabian Style

Bonilla, Diego A., Isabel A. Sánchez-Rojas, Darío Mendoza-Romero, Yurany Moreno, Jana Kočí, Luis M. Gómez-Miranda, Daniel Rojas-Valverde, Jorge L. Petro, and Richard B. Kreider. 2023. "Profiling Physical Fitness of Physical Education Majors Using Unsupervised Machine Learning" International Journal of Environmental Research and Public Health 20, no. 1: 146. https://doi.org/10.3390/ijerph20010146

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop