Panagiotis Aggelakis
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Panagiotis Aggelakis has a total of 49 co-authors. Here are the top co-authors by number of shared publications:
- Alexandra Avloniti
Department of Physical Education and Sport Science, School of Physical Education, Sport Science and Occupational Therapy, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications - Athanasios Chatzinikolaou
Department of Physical Education and Sport Science, School of Physical Education, Sport Science and Occupational Therapy, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications - Christos Kokkotis
Department of Occupational Therapy, School of Physical Education, Sport Science and Occupational Therapy, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications - Dimitrios Balampanos
Department of Physical Education and Sport Science, School of Physical Education, Sport Science and Occupational Therapy, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications - Dimitrios Pantazis
Department of Physical Education and Sport Science, School of Physical Education, Sport Science and Occupational Therapy, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications - Nikolaos Zaras
Department of Physical Education and Sport Science, Democritus University of Thrace, 69100 Komotini, Greece
15 shared publications
Latest Publications
- Open Access
- 21 July 2026
- Published by MDPI AG in Journal of Functional Morphology and Kinesiology
- Vol. 11  (3) , 282
Performance Profiles in Youth Basketball Across Different Score Contexts: An Unsupervised Machine Learning Analysis
Objectives: The analysis of basketball performance has increasingly incorporated advanced analytics and machine learning methods to better understand the factors that influence offensive efficiency and match dynamics. The present study aimed to identify performance profiles in basketball using unsupervised machine learning techniques and to examine the physical load and performance indicators that differentiate these profiles. Methods: Team-quarter observations from the Final 8 phase of the Greek U16 Basketball Championship were stratified into quarters with large score differences and quarters with small score differences according to the quarter-specific score differential and the sample median of 5 points. K-means clustering was applied separately to each dataset to identify latent performance patterns. Candidate solutions were evaluated using the Elbow method, Silhouette coefficient, Calinski–Harabasz Index, and Davies–Bouldin Index. Based on their combined interpretation, together with considerations of parsimony and practical interpretability, two-cluster solutions were retained for both datasets. Cluster stability was assessed using the Adjusted Rand Index (ARI), while t-distributed stochastic neighbor embedding (t-SNE) was used exclusively for visualization of the identified clusters. Results: Welch’s independent-samples t-tests with Benjamini–Hochberg false discovery rate (FDR) correction identified significant differences between clusters across several external load variables, including jump load, total distance covered, accumulated acceleration load, and distance covered in different speed zones (pFDR < 0.001). Clusters characterized by higher movement intensity also exhibited higher values for basketball performance and offensive-efficiency indicators. Although higher-performance clusters showed numerically higher winning proportions in both contexts (large score differences: 70.0% vs. 45.7%; small score differences: 56.7% vs. 40.6%), chi-square analyses indicated that cluster membership was not significantly associated with quarter outcomes. Conclusions: Overall, the findings suggest that performance profiles in basketball are primarily differentiated by external-load characteristics, particularly movement intensity, and offensive-performance indicators, highlighting the importance of integrating both physical and technical performance indicators in basketball performance analysis.
- 21 July 2026
- Published by MDPI AG in Journal of Functional Morphology and Kinesiology
- Vol. 11  (3) , 282
Open AccessPerformance Profiles in Youth Basketball Across Different Score Contexts: An Unsupervised Machine Learning Analysis
Objectives: The analysis of basketball performance has increasingly incorporated advanced analytics and machine learning methods to better understand the factors that influence offensive efficiency and match dynamics. The present study aimed to identify performance profiles in basketball using unsupervised machine learning techniques and to examine the physical load and performance indicators that differentiate these profiles. Methods: Team-quarter observations from the Final 8 phase of the Greek U16 Basketball Championship were stratified into quarters with large score differences and quarters with small score differences according to the quarter-specific score differential and the sample median of 5 points. K-means clustering was applied separately to each dataset to identify latent performance patterns. Candidate solutions were evaluated using the Elbow method, Silhouette coefficient, Calinski–Harabasz Index, and Davies–Bouldin Index. Based on their combined interpretation, together with considerations of parsimony and practical interpretability, two-cluster solutions were retained for both datasets. Cluster stability was assessed using the Adjusted Rand Index (ARI), while t-distributed stochastic neighbor embedding (t-SNE) was used exclusively for visualization of the identified clusters. Results: Welch’s independent-samples t-tests with Benjamini–Hochberg false discovery rate (FDR) correction identified significant differences between clusters across several external load variables, including jump load, total distance covered, accumulated acceleration load, and distance covered in different speed zones (pFDR < 0.001). Clusters characterized by higher movement intensity also exhibited higher values for basketball performance and offensive-efficiency indicators. Although higher-performance clusters showed numerically higher winning proportions in both contexts (large score differences: 70.0% vs. 45.7%; small score differences: 56.7% vs. 40.6%), chi-square analyses indicated that cluster membership was not significantly associated with quarter outcomes. Conclusions: Overall, the findings suggest that performance profiles in basketball are primarily differentiated by external-load characteristics, particularly movement intensity, and offensive-performance indicators, highlighting the importance of integrating both physical and technical performance indicators in basketball performance analysis.
- Open Access
- 25 June 2026
- Published by MDPI AG in Journal of Functional Morphology and Kinesiology
- Vol. 11  (3) , 249
Data-Driven Reduction of External Load Variables in Indoor Team Sports Using Local Positioning System
Objectives: Local positioning systems (LPSs) used in indoor team sports generate a large number of external load variables, often exceeding practical monitoring capacity. The redundancy and overlap among these variables make it difficult to identify the most informative metrics for performance analysis and load management. This study aimed to reduce the dimensionality of external load variables derived from LPS data and to identify data-driven external-load observation profiles using principal component analysis and clustering techniques. Methods: A total of 188 observations from indoor team sports (basketball, handball, and futsal) were analyzed. Continuous external load variables were standardized and subjected to principal component analysis (PCA), with component retention based on a ≥90% cumulative explained variance threshold. K-means clustering was applied in both the full standardized feature space and the PCA-reduced space. The optimal number of clusters was determined using silhouette analysis and the elbow method. Agreement between clustering solutions was assessed using Adjusted Rand Index (ARI) and Normalized Mutual Information (NMI). Cluster characteristics were further examined using descriptive statistics and variable separation analysis. Results: The first two principal components explained 53.7% of the total variance, representing high-intensity external load and neuromuscular load dimensions, while 12 components were required to exceed 90% cumulative explained variance. Clustering analysis consistently identified three moderately separated clusters in both the full and PCA-reduced spaces. The PCA-based solution demonstrated improved separation (silhouette = 0.362) compared to the full-space solution (silhouette = 0.319). Agreement between clustering approaches was high (ARI = 0.981; NMI = 0.971), indicating that dimensionality reduction largely preserved the main clustering structure within the analyzed dataset. The most discriminative variables included jump load, acceleration load, metabolic power, and anaerobic activity distance. Conclusions: A large set of external load variables can be reduced into interpretable latent dimensions that support exploratory external-load profile identification. The combination of PCA and clustering provides an exploratory and structure-preserving framework for summarizing complex external-load datasets and identifying latent load dimensions. These findings may assist future monitoring strategies; however, the practical utility of the identified profiles requires prospective validation before implementation in training-load management.
- 25 June 2026
- Published by MDPI AG in Journal of Functional Morphology and Kinesiology
- Vol. 11  (3) , 249
Open AccessData-Driven Reduction of External Load Variables in Indoor Team Sports Using Local Positioning System
Objectives: Local positioning systems (LPSs) used in indoor team sports generate a large number of external load variables, often exceeding practical monitoring capacity. The redundancy and overlap among these variables make it difficult to identify the most informative metrics for performance analysis and load management. This study aimed to reduce the dimensionality of external load variables derived from LPS data and to identify data-driven external-load observation profiles using principal component analysis and clustering techniques. Methods: A total of 188 observations from indoor team sports (basketball, handball, and futsal) were analyzed. Continuous external load variables were standardized and subjected to principal component analysis (PCA), with component retention based on a ≥90% cumulative explained variance threshold. K-means clustering was applied in both the full standardized feature space and the PCA-reduced space. The optimal number of clusters was determined using silhouette analysis and the elbow method. Agreement between clustering solutions was assessed using Adjusted Rand Index (ARI) and Normalized Mutual Information (NMI). Cluster characteristics were further examined using descriptive statistics and variable separation analysis. Results: The first two principal components explained 53.7% of the total variance, representing high-intensity external load and neuromuscular load dimensions, while 12 components were required to exceed 90% cumulative explained variance. Clustering analysis consistently identified three moderately separated clusters in both the full and PCA-reduced spaces. The PCA-based solution demonstrated improved separation (silhouette = 0.362) compared to the full-space solution (silhouette = 0.319). Agreement between clustering approaches was high (ARI = 0.981; NMI = 0.971), indicating that dimensionality reduction largely preserved the main clustering structure within the analyzed dataset. The most discriminative variables included jump load, acceleration load, metabolic power, and anaerobic activity distance. Conclusions: A large set of external load variables can be reduced into interpretable latent dimensions that support exploratory external-load profile identification. The combination of PCA and clustering provides an exploratory and structure-preserving framework for summarizing complex external-load datasets and identifying latent load dimensions. These findings may assist future monitoring strategies; however, the practical utility of the identified profiles requires prospective validation before implementation in training-load management.
- Open Access
- 10 June 2026
- Published by MDPI AG in Preprints.org
Data-Driven Reduction of External Load Variables in Indoor Team Sports Using Local Positioning System
Objectives: Local positioning systems (LPS) used in indoor team sports generate a large number of external load variables, often exceeding practical monitoring capacity. The redundancy and overlap among these variables make it difficult to identify the most informative metrics for performance analysis and load management. This study aimed to reduce the dimensionality of external load variables derived from LPS data and to identify data-driven external-load observation profiles using principal component analysis and clustering techniques. Methods: A total of 188 observations from indoor team sports (basketball, handball, and futsal) were analyzed. Continuous external load variables were standardized and subjected to principal component analysis (PCA), with component retention based on a ≥90% cumulative explained variance threshold. K-means clustering was applied in both the full standardized feature space and the PCA-reduced space. The optimal number of clusters was determined using silhouette analysis and the elbow method. Agreement between clustering solutions was assessed using Adjusted Rand Index (ARI) and Normalized Mutual Information (NMI). Cluster characteristics were further examined using descriptive statistics and variable separation analysis. Results: The first two principal components explained 53.7% of the total variance, representing high-intensity external load and neuromuscular load dimensions, while 12 components were required to exceed 90% cumulative explained variance. Clustering analysis consistently identified three moderately separated clusters in both the full and PCA-reduced spaces. The PCA-based solution demonstrated improved separation (silhouette = 0.362) compared to the full-space solution (silhouette = 0.319). Agreement between clustering approaches was high (ARI = 0.981; NMI = 0.971), indicating that dimensionality reduction largely preserved the main clustering structure within the analyzed dataset. The most discriminative variables included jump load, acceleration load, metabolic power, and anaerobic activity distance. Conclusions: A large set of external load variables can be reduced into interpretable latent dimensions that supported exploratory external-load profile identification. The combination of PCA and clustering provides an exploratory, structure-preserving, and interpretable framework for simplifying load monitoring in indoor team sports, supporting the selection of key performance indicators and more efficient training load management.
- 10 June 2026
- Published by MDPI AG in Preprints.org
Open AccessData-Driven Reduction of External Load Variables in Indoor Team Sports Using Local Positioning System
Objectives: Local positioning systems (LPS) used in indoor team sports generate a large number of external load variables, often exceeding practical monitoring capacity. The redundancy and overlap among these variables make it difficult to identify the most informative metrics for performance analysis and load management. This study aimed to reduce the dimensionality of external load variables derived from LPS data and to identify data-driven external-load observation profiles using principal component analysis and clustering techniques. Methods: A total of 188 observations from indoor team sports (basketball, handball, and futsal) were analyzed. Continuous external load variables were standardized and subjected to principal component analysis (PCA), with component retention based on a ≥90% cumulative explained variance threshold. K-means clustering was applied in both the full standardized feature space and the PCA-reduced space. The optimal number of clusters was determined using silhouette analysis and the elbow method. Agreement between clustering solutions was assessed using Adjusted Rand Index (ARI) and Normalized Mutual Information (NMI). Cluster characteristics were further examined using descriptive statistics and variable separation analysis. Results: The first two principal components explained 53.7% of the total variance, representing high-intensity external load and neuromuscular load dimensions, while 12 components were required to exceed 90% cumulative explained variance. Clustering analysis consistently identified three moderately separated clusters in both the full and PCA-reduced spaces. The PCA-based solution demonstrated improved separation (silhouette = 0.362) compared to the full-space solution (silhouette = 0.319). Agreement between clustering approaches was high (ARI = 0.981; NMI = 0.971), indicating that dimensionality reduction largely preserved the main clustering structure within the analyzed dataset. The most discriminative variables included jump load, acceleration load, metabolic power, and anaerobic activity distance. Conclusions: A large set of external load variables can be reduced into interpretable latent dimensions that supported exploratory external-load profile identification. The combination of PCA and clustering provides an exploratory, structure-preserving, and interpretable framework for simplifying load monitoring in indoor team sports, supporting the selection of key performance indicators and more efficient training load management.
- Open Access
- 10 June 2026
- Published by MDPI AG in Preprints.org
Performance Profiles in Youth Basketball Across Different Score Contexts: An Unsupervised Machine Learning Analysis
The analysis of basketball performance has increasingly incorporated advanced analytics and machine learning methods to understand better the factors that influence offensive efficiency and game dynamics. The present study aimed to identify performance profiles in basketball using unsupervised machine learning techniques and to examine the physical load and performance indicators that differentiate these profiles. Quarters from Greek U16 basketball matches were analyzed in two contexts: quarters from games with large score differences and quarters from games with small score differences. K-means clustering was applied separately to each dataset to identify latent performance patterns. The optimal number of clusters was determined using the Elbow method and silhouette analysis, which indicated a two-cluster solution for both datasets. Cluster visualization using t-distributed stochastic neighbor embedding (t-SNE) confirmed the presence of distinct performance profiles. Independent samples t-tests revealed significant differences between clusters across several physical load indicators, including jump load, total distance covered, accumulated acceleration load, and distance covered across different speed zones (p < .001). Clusters characterized by higher movement intensity also exhibited higher basketball performance efficiency indicators. Although higher-performance clusters showed numerically higher winning proportions in both contexts (large score differences: 66.7% vs. 42.9%; small score differences: 59.4% vs. 36.7%), chi-square analyses indicated that cluster membership was not significantly associated with game outcomes. Overall, the findings suggest that performance profiles in basketball are primarily differentiated by movement intensity and physical load characteristics, highlighting the importance of integrating both physical and technical performance indicators in basketball performance analysis.
- 10 June 2026
- Published by MDPI AG in Preprints.org
Open AccessPerformance Profiles in Youth Basketball Across Different Score Contexts: An Unsupervised Machine Learning Analysis
The analysis of basketball performance has increasingly incorporated advanced analytics and machine learning methods to understand better the factors that influence offensive efficiency and game dynamics. The present study aimed to identify performance profiles in basketball using unsupervised machine learning techniques and to examine the physical load and performance indicators that differentiate these profiles. Quarters from Greek U16 basketball matches were analyzed in two contexts: quarters from games with large score differences and quarters from games with small score differences. K-means clustering was applied separately to each dataset to identify latent performance patterns. The optimal number of clusters was determined using the Elbow method and silhouette analysis, which indicated a two-cluster solution for both datasets. Cluster visualization using t-distributed stochastic neighbor embedding (t-SNE) confirmed the presence of distinct performance profiles. Independent samples t-tests revealed significant differences between clusters across several physical load indicators, including jump load, total distance covered, accumulated acceleration load, and distance covered across different speed zones (p < .001). Clusters characterized by higher movement intensity also exhibited higher basketball performance efficiency indicators. Although higher-performance clusters showed numerically higher winning proportions in both contexts (large score differences: 66.7% vs. 42.9%; small score differences: 59.4% vs. 36.7%), chi-square analyses indicated that cluster membership was not significantly associated with game outcomes. Overall, the findings suggest that performance profiles in basketball are primarily differentiated by movement intensity and physical load characteristics, highlighting the importance of integrating both physical and technical performance indicators in basketball performance analysis.
- Open Access
- 7 June 2026
- Published by MDPI AG in Algorithms
- Vol. 19  (6) , 464
Proof of Concept for a Deep-Learning Computer-Vision System to Quantify External Load in Basketball: Comparison with Local Positioning Systems
Background: Monitoring external load in team sports is essential for performance optimization, injury prevention, and individualized training prescription. Although Local Positioning Systems (LPS) are widely used for indoor athlete tracking, they require wearable devices and specialized infrastructure. Recent advances in artificial intelligence and computer vision allow markerless athlete tracking; however, their validity for basketball remains insufficiently explored. Objective: To evaluate the validity of a deep-learning multi-camera computer-vision system for quantifying external-load variables in basketball compared with a commercial LPS. Methods: The framework integrated fisheye video acquisition, player detection, and pose estimation using YOLOv11x-Pose and player re-identification through ResNet-50 and FAISS similarity search. Positional data were transformed into real-world court coordinates to derive distance, acceleration, deceleration, player load, and average speed metrics. Outputs were compared with measurements obtained from Kinexon LPS. Results: Strong correlations were observed for total distance (r = 0.92), acceleration counts (r = 0.90), deceleration counts (r = 0.92), and player load (r = 0.81), while average speed showed a moderate-to-strong correlation (r = 0.66). ICC and Bland–Altman analyses indicated agreement between systems. Conclusions: The proposed computer-vision system demonstrated high agreement with LPS, supporting its use as a valid, non-invasive, and scalable solution for external load monitoring in basketball.
- 7 June 2026
- Published by MDPI AG in Algorithms
- Vol. 19  (6) , 464
Open AccessProof of Concept for a Deep-Learning Computer-Vision System to Quantify External Load in Basketball: Comparison with Local Positioning Systems
Background: Monitoring external load in team sports is essential for performance optimization, injury prevention, and individualized training prescription. Although Local Positioning Systems (LPS) are widely used for indoor athlete tracking, they require wearable devices and specialized infrastructure. Recent advances in artificial intelligence and computer vision allow markerless athlete tracking; however, their validity for basketball remains insufficiently explored. Objective: To evaluate the validity of a deep-learning multi-camera computer-vision system for quantifying external-load variables in basketball compared with a commercial LPS. Methods: The framework integrated fisheye video acquisition, player detection, and pose estimation using YOLOv11x-Pose and player re-identification through ResNet-50 and FAISS similarity search. Positional data were transformed into real-world court coordinates to derive distance, acceleration, deceleration, player load, and average speed metrics. Outputs were compared with measurements obtained from Kinexon LPS. Results: Strong correlations were observed for total distance (r = 0.92), acceleration counts (r = 0.90), deceleration counts (r = 0.92), and player load (r = 0.81), while average speed showed a moderate-to-strong correlation (r = 0.66). ICC and Bland–Altman analyses indicated agreement between systems. Conclusions: The proposed computer-vision system demonstrated high agreement with LPS, supporting its use as a valid, non-invasive, and scalable solution for external load monitoring in basketball.
- Open Access
Physical Activity During Official Match Play in Female Masters Basketball Players: An Accelerometry-Based Study
Background/Objectives: Insufficient physical activity remains a major public health concern among adult women, highlighting the need to identify structured activity contexts that can contribute meaningfully to recommended weekly physical activity levels. Official masters basketball may represent one such context; however, the amount of physical activity accumulated during female masters basketball match play remains insufficiently quantified. This study quantified the physical activity profile of official tournament match play among female masters basketball athletes and described the associated external physical demands. Methods: This observational study included 52 female master basketball athletes aged 37–63 years who competed in a three-day national masters tournament. Match demands were monitored using tri-axial microsensors. Physical activity was classified from processed raw tri-axial acceleration data into intensity zones, and differences in time spent across zones were examined using one-way repeated-measures ANOVA. External load during active play was quantified using total distance, distance across speed zones, accumulated acceleration load (AAL), mechanical load (ML), jump load (JL), and Physio Load. Results: Significant differences were observed across physical-activity intensity zones, with more time accumulated in light physical activity (LPA) and vigorous physical activity (VPA) than in moderate physical activity (MPA), whereas MPA accounted for the least time overall [F (1.98, 101.16) = 47.57, p < 0.001, ηp2 = 0.48]. Descriptively, moderate-to-vigorous physical activity (MVPA) amounted to 42.78 min, calculated as the sum of MPA (9.41 ± 3.82 min) and VPA (33.37 ± 14.49 min). During active play, athletes covered 59.19 ± 17.26 m·min−1, with most distance accumulated in the low- and medium-speed zones and limited very-high-speed running; AAL, ML, and JL averaged 8.32 ± 2.31 AU·min−1, 22.35 ± 5.53 AU·min−1, and 31.26 ± 28.35 J·min−1, respectively. Conclusions: Official female masters basketball appears to provide a meaningful intermittent physical-activity stimulus within a single monitored match exposure and may contribute substantially to weekly aerobic physical-activity accumulation in adult women.
Open AccessPhysical Activity During Official Match Play in Female Masters Basketball Players: An Accelerometry-Based Study
Background/Objectives: Insufficient physical activity remains a major public health concern among adult women, highlighting the need to identify structured activity contexts that can contribute meaningfully to recommended weekly physical activity levels. Official masters basketball may represent one such context; however, the amount of physical activity accumulated during female masters basketball match play remains insufficiently quantified. This study quantified the physical activity profile of official tournament match play among female masters basketball athletes and described the associated external physical demands. Methods: This observational study included 52 female master basketball athletes aged 37–63 years who competed in a three-day national masters tournament. Match demands were monitored using tri-axial microsensors. Physical activity was classified from processed raw tri-axial acceleration data into intensity zones, and differences in time spent across zones were examined using one-way repeated-measures ANOVA. External load during active play was quantified using total distance, distance across speed zones, accumulated acceleration load (AAL), mechanical load (ML), jump load (JL), and Physio Load. Results: Significant differences were observed across physical-activity intensity zones, with more time accumulated in light physical activity (LPA) and vigorous physical activity (VPA) than in moderate physical activity (MPA), whereas MPA accounted for the least time overall [F (1.98, 101.16) = 47.57, p < 0.001, ηp2 = 0.48]. Descriptively, moderate-to-vigorous physical activity (MVPA) amounted to 42.78 min, calculated as the sum of MPA (9.41 ± 3.82 min) and VPA (33.37 ± 14.49 min). During active play, athletes covered 59.19 ± 17.26 m·min−1, with most distance accumulated in the low- and medium-speed zones and limited very-high-speed running; AAL, ML, and JL averaged 8.32 ± 2.31 AU·min−1, 22.35 ± 5.53 AU·min−1, and 31.26 ± 28.35 J·min−1, respectively. Conclusions: Official female masters basketball appears to provide a meaningful intermittent physical-activity stimulus within a single monitored match exposure and may contribute substantially to weekly aerobic physical-activity accumulation in adult women.
- Open Access
- 18 April 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16  (8) , 3951
Does Basketball Training Load Provide an Adequate Amount of Physical Activity for Pre-Peak Height Velocity Athletes?
Purpose: The primary aim of the present study was to examine the extent to which participation in organized youth basketball training contributes to physical activity across intensity zones during training sessions in relation to biological maturation status. Methods: Participants were classified into three maturity groups based on predicted age at peak height velocity (PHV): −2.5 to −1.5, −1.5 to −0.5, and ≥−0.5 to 0.83 years from PHV. Data from two training sessions per participant were averaged to obtain representative individual values. One-way analyses of variance (ANOVAs) were used to examine differences in anthropometric, physical performance, and field performance variables between PHV groups. Physical activity patterns were analyzed using two-way mixed-design ANOVAs with PHV stage as the between-subject factor and intensity zone (MET- and HRR-based) as the within-subject factor. Results: Across all maturity groups, approximately 10–17% of total training time was spent in light-intensity activity, while the majority of time was accumulated in moderate-to-vigorous intensity zones (approximately 35–50%, depending on the classification method). Significant maturity-related differences were observed in anthropometric variables and physical performance measures, with more mature players demonstrating superior sprint performance, jumping ability, and grip strength. Field performance indicators also differed between PHV groups, with more mature athletes exhibiting higher external and internal training loads. In contrast, no significant interactions or main effects of PHV stage were observed for physical activity intensity distribution. Conclusions: Organized basketball training contributes substantially to moderate-to-vigorous physical activity accumulated during training sessions. However, these findings reflect training-specific activity and should not be interpreted as representing total daily physical activity. No differences in activity intensity distribution were observed between maturation groups, although this finding should be interpreted with caution, given methodological limitations. These results highlight the need to consider biological maturation when designing youth training programs.
- 18 April 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16  (8) , 3951
Open AccessDoes Basketball Training Load Provide an Adequate Amount of Physical Activity for Pre-Peak Height Velocity Athletes?
Purpose: The primary aim of the present study was to examine the extent to which participation in organized youth basketball training contributes to physical activity across intensity zones during training sessions in relation to biological maturation status. Methods: Participants were classified into three maturity groups based on predicted age at peak height velocity (PHV): −2.5 to −1.5, −1.5 to −0.5, and ≥−0.5 to 0.83 years from PHV. Data from two training sessions per participant were averaged to obtain representative individual values. One-way analyses of variance (ANOVAs) were used to examine differences in anthropometric, physical performance, and field performance variables between PHV groups. Physical activity patterns were analyzed using two-way mixed-design ANOVAs with PHV stage as the between-subject factor and intensity zone (MET- and HRR-based) as the within-subject factor. Results: Across all maturity groups, approximately 10–17% of total training time was spent in light-intensity activity, while the majority of time was accumulated in moderate-to-vigorous intensity zones (approximately 35–50%, depending on the classification method). Significant maturity-related differences were observed in anthropometric variables and physical performance measures, with more mature players demonstrating superior sprint performance, jumping ability, and grip strength. Field performance indicators also differed between PHV groups, with more mature athletes exhibiting higher external and internal training loads. In contrast, no significant interactions or main effects of PHV stage were observed for physical activity intensity distribution. Conclusions: Organized basketball training contributes substantially to moderate-to-vigorous physical activity accumulated during training sessions. However, these findings reflect training-specific activity and should not be interpreted as representing total daily physical activity. No differences in activity intensity distribution were observed between maturation groups, although this finding should be interpreted with caution, given methodological limitations. These results highlight the need to consider biological maturation when designing youth training programs.
- Open Access
- 19 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16  (4) , 2037
Individualized Physical Performance Metrics in 3 × 3 Basketball Games Using Match-Play Data
3 × 3 basketball is a high-intensity intermittent sport practiced by both professional and recreational athletes. However, the use of predefined absolute thresholds to quantify external load may overlook meaningful inter-individual differences in movement intensity. This study examined internal and external load demands during official 3 × 3 match play using individualized, performance-based load zones. Seventeen male players were monitored across 38 valid match observations during a two-day tournament. External load was collected via inertial measurement units, while internal load was assessed through continuous heart-rate monitoring. Raw triaxial accelerometer data were processed in Python to remove gravitational components and reconstruct speed–acceleration profiles, allowing identification of individual acceleration, deceleration, and jump events. Statistical analyses were conducted using linear mixed-effects models with Bonferroni-adjusted post hoc comparisons to evaluate differences between absolute and individualized zones. Players sustained high physiological strain, operating at approximately 85–90% of HRmax, and performed frequent high-intensity mechanical actions. Individualized acceleration, deceleration, and jump zones yielded a more even dispersion of events across low-, moderate-, and high-intensity categories. In contrast, predefined absolute thresholds classified over 90% of events as low intensity, masking meaningful variability. These findings highlight substantial inter-individual differences in 3 × 3 match demands and support the use of individualized load profiling for accurate monitoring, performance evaluation, and training prescription.
- 19 February 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16  (4) , 2037
Open AccessIndividualized Physical Performance Metrics in 3 × 3 Basketball Games Using Match-Play Data
3 × 3 basketball is a high-intensity intermittent sport practiced by both professional and recreational athletes. However, the use of predefined absolute thresholds to quantify external load may overlook meaningful inter-individual differences in movement intensity. This study examined internal and external load demands during official 3 × 3 match play using individualized, performance-based load zones. Seventeen male players were monitored across 38 valid match observations during a two-day tournament. External load was collected via inertial measurement units, while internal load was assessed through continuous heart-rate monitoring. Raw triaxial accelerometer data were processed in Python to remove gravitational components and reconstruct speed–acceleration profiles, allowing identification of individual acceleration, deceleration, and jump events. Statistical analyses were conducted using linear mixed-effects models with Bonferroni-adjusted post hoc comparisons to evaluate differences between absolute and individualized zones. Players sustained high physiological strain, operating at approximately 85–90% of HRmax, and performed frequent high-intensity mechanical actions. Individualized acceleration, deceleration, and jump zones yielded a more even dispersion of events across low-, moderate-, and high-intensity categories. In contrast, predefined absolute thresholds classified over 90% of events as low intensity, masking meaningful variability. These findings highlight substantial inter-individual differences in 3 × 3 match demands and support the use of individualized load profiling for accurate monitoring, performance evaluation, and training prescription.
- Open Access
- 21 November 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15  (23) , 12401
From Data to Decisions: Using Explainable Machine Learning to Predict EuroLeague Basketball Outcomes
Predicting basketball game outcomes in elite competitions is a complex task influenced by multiple interacting performance factors. This study applied a supervised machine learning (ML) framework to predict EuroLeague game outcomes using team-level game-related statistics. Four algorithms—Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and Naïve Bayes (NB)—were trained and compared following recursive feature elimination (RFE) to identify the most informative predictors. The dataset comprised comprehensive in-game statistics describing shooting efficiency, rebounding, ball security, and spatial shot distribution. Model performance was evaluated using accuracy, area under the receiver operating characteristic curve (AUC), precision, recall, and F1-score, ensuring both discrimination and calibration assessment. Among the four classifiers, SVM (AUC = 0.922, Accuracy = 0.841) and LR (AUC = 0.933, Accuracy = 0.818) achieved the highest predictive performance, outperforming RF and NB. Feature importance analysis using Shapley Additive Explanations (SHAP) on the best-performing SVM classifier revealed that true shooting percentage (TS%), defensive rebounds (DR), steals (ST), and turnovers (TO) were the most influential predictors of game outcomes. Teams that demonstrated higher shooting efficiency, greater rebounding control, and fewer turnovers showed a significantly higher probability of winning. These results confirm that well-validated and interpretable ML models can accurately predict game outcomes in professional basketball using readily available box-score statistics. The integration of RFE-based feature selection and SHAP interpretability provides transparent, evidence-based insights that can inform tactical decisions, enhance scouting accuracy, and support coaches in developing data-driven performance strategies within elite basketball environments.
- 21 November 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15  (23) , 12401
Open AccessFrom Data to Decisions: Using Explainable Machine Learning to Predict EuroLeague Basketball Outcomes
Predicting basketball game outcomes in elite competitions is a complex task influenced by multiple interacting performance factors. This study applied a supervised machine learning (ML) framework to predict EuroLeague game outcomes using team-level game-related statistics. Four algorithms—Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and Naïve Bayes (NB)—were trained and compared following recursive feature elimination (RFE) to identify the most informative predictors. The dataset comprised comprehensive in-game statistics describing shooting efficiency, rebounding, ball security, and spatial shot distribution. Model performance was evaluated using accuracy, area under the receiver operating characteristic curve (AUC), precision, recall, and F1-score, ensuring both discrimination and calibration assessment. Among the four classifiers, SVM (AUC = 0.922, Accuracy = 0.841) and LR (AUC = 0.933, Accuracy = 0.818) achieved the highest predictive performance, outperforming RF and NB. Feature importance analysis using Shapley Additive Explanations (SHAP) on the best-performing SVM classifier revealed that true shooting percentage (TS%), defensive rebounds (DR), steals (ST), and turnovers (TO) were the most influential predictors of game outcomes. Teams that demonstrated higher shooting efficiency, greater rebounding control, and fewer turnovers showed a significantly higher probability of winning. These results confirm that well-validated and interpretable ML models can accurately predict game outcomes in professional basketball using readily available box-score statistics. The integration of RFE-based feature selection and SHAP interpretability provides transparent, evidence-based insights that can inform tactical decisions, enhance scouting accuracy, and support coaches in developing data-driven performance strategies within elite basketball environments.
- Open Access
Exploring Bone Health Determinants in Youth Athletes Using Supervised and Unsupervised Machine Learning
Background: Bone health in youth is influenced by both modifiable factors, such as nutrition and physical activity, and non-modifiable factors, such as biological maturation and heredity. Understanding how these elements interact to predict body composition may enhance the effectiveness of early interventions. Importantly, the integration of both supervised and unsupervised machine learning models enables a data-driven exploration of complex relationships, allowing for accurate prediction and subgroup discovery. Methods: This cross-sectional study examined 94 male athletes during the developmental period. Anthropometric, performance, and nutritional data were collected, and bone parameters were assessed using dual-energy X-ray absorptiometry (DXA). Three supervised machine learning models (Random Forest, Gradient Boosting, and Support Vector Regression) were trained to predict Total Body-Less Head (TBLH) values. Nested cross-validation assessed model performance. Unsupervised clustering (K-Means) was also applied to identify dietary intake profiles (calcium, protein, vitamin D). SHAP analysis was used for model interpretability. Results: The Random Forest model yielded the best predictive performance (R2 = 0.71, RMSE = 0.057). Weight, height, and handgrip strength were the most influential predictors. Clustering analysis revealed two distinct nutritional profiles; however, t-tests showed no significant differences in TBLH or regional BMD between the clusters. Conclusions: Machine learning, both supervised for accurate prediction and unsupervised for nutritional subgroup discovery, provides a robust, interpretable framework for assessing adolescent bone health. While dietary intake clusters did not align with significant differences in bone parameters, this finding underscores the multifactorial nature of skeletal development and highlights areas for further exploration.
Open AccessExploring Bone Health Determinants in Youth Athletes Using Supervised and Unsupervised Machine Learning
Background: Bone health in youth is influenced by both modifiable factors, such as nutrition and physical activity, and non-modifiable factors, such as biological maturation and heredity. Understanding how these elements interact to predict body composition may enhance the effectiveness of early interventions. Importantly, the integration of both supervised and unsupervised machine learning models enables a data-driven exploration of complex relationships, allowing for accurate prediction and subgroup discovery. Methods: This cross-sectional study examined 94 male athletes during the developmental period. Anthropometric, performance, and nutritional data were collected, and bone parameters were assessed using dual-energy X-ray absorptiometry (DXA). Three supervised machine learning models (Random Forest, Gradient Boosting, and Support Vector Regression) were trained to predict Total Body-Less Head (TBLH) values. Nested cross-validation assessed model performance. Unsupervised clustering (K-Means) was also applied to identify dietary intake profiles (calcium, protein, vitamin D). SHAP analysis was used for model interpretability. Results: The Random Forest model yielded the best predictive performance (R2 = 0.71, RMSE = 0.057). Weight, height, and handgrip strength were the most influential predictors. Clustering analysis revealed two distinct nutritional profiles; however, t-tests showed no significant differences in TBLH or regional BMD between the clusters. Conclusions: Machine learning, both supervised for accurate prediction and unsupervised for nutritional subgroup discovery, provides a robust, interpretable framework for assessing adolescent bone health. While dietary intake clusters did not align with significant differences in bone parameters, this finding underscores the multifactorial nature of skeletal development and highlights areas for further exploration.