Table 2. Data analysis characteristics.
Author (Year) Lower Extremity Location Predictor Variables Data Pre- processing Feature Selection/ Dimensionality Reduction Training Strategy
Lopez-Valenciano et al. (2018) Any muscle (traumatic) - Demographics and Injury History (Age, BMI, Injury history, Sleep, Level of play)
- Psychological and Perceptual Variables
(Sport devaluation, Sleep quality)
- Physical Performance Measures (YBalance test, Core control, ROM of hip, Isometric strength)
-Data imputation
-Weka software
NR - SMOTE
-5-fold cross-validation
Ruddy et al. (2018) HSI (traumatic) - Demographics and Injury History (Age, Height, Mass. Playing Position, History HIS, History ACL)
- Physical Performance Measures (Peak Hamstring Force Right, Peak Hamstring Force Left, Hamstring Force Imbalance)
Data normalized (Z-score) NR - SMOTE
- 10-fold cross-validation
Ayala et al. (2019) HSI (traumatic) - Demographics and Injury History (Age, History of HSI last season, Maximal level of play achieved)
- Psychological and Perceptual Variables (Sleep quality, Physical/emotional exhaustion, Reduced sense of accomplishment)
- Physical Performance Measures (Dynamic postural control, Isometric hip abduction and adduction strength, Lower extremity joint ROMs)
Data imputation NR - SMOTE
- 3-fold cross-validation
Connaboy et al. (2019) Any region (NR) - Demographics (Age, Boday fat, Weight)
- Physical Performance Measures (Peak anaerobic power, Mean anaerobic power, Knee active extension)
Not Reported NR Leave-one out cross validation
Henriquez et al. (2020) Any region (NR) - Demographics (Weight, Height, Gender, Age)
- Physical Performance Measures (Eyes Open Balance Test Composite Score, DPSI Composite Score, Straight Leg Raise, Active Knee Extension, Ankle Dorsiflexion Strength)
Data normalized (Z-score) Mean Decrease Accuracy 5-fold cross-validation
Oliver et al. (2020) Any region (traumatic) -Demographic (Age, BMI, Height);
-Physical Performance Measures (Maturity-Offset, 75%Hop L PVGRF, 75%Hop R PVGRF)
Weka software NR - Cost-sensitive learning
- 5-fold
cross-validation
Jauhiainen et al. (2021) Knee, Ankle (traumatic) - Demographic and Injury History (Age, Sex, BMI, Previous ACL, Family ACL history)
- Physical Performance Measures (KT1000 (dominant leg), hip flexion peak (dominant leg), medial knee displacement (both legs), vertical ground reaction force (vGRF) (both legs))
-Data imputation
-Data normalized (Z-score)
Expert-based feature selection 10-fold cross-validation
Ruiz-Perez et al. (2021) Any region (traumatic) - Demographics and Injury History (Player position, Current level of play, Dominant leg, Sex, Age)
- Psychological and Perceptual Variables (Physical/emotional exhaustion, Sleep quality)
- Physical Performance Measures (PosteroLateral, Y-Balance-Composite)
-Data imputation
-Weka software
Attribute Selected Classifier -Under-sampling Bagging
-5-fold cross-validation
Bogaert et al. (2022) Any region (overuse) - Demographics and Injury History (Gender, Weight, Height, Previous injuries)
- Physical Performance Measures (Root-mean-square ratio, Step regularity, Stride regularity, Sample entropy)
Data normalized (Min-Max Scaling) PCA (Principal Component Analysis) -Cost-sensitive learning
-Internal Cross-Validation
Jauhiainen et al. (2022) Knee (traumatic) - Demographics and Injury History (Age, Body mass, Previous ACL)
- Physical Performance Measures (Single leg drop jump knee, Jump hip flex max)
-Data imputation
-Data normalized
NR - SMOTE
-5-fold cross-validation
Huang et al. (2022) Any region (overuse) - Psychological and Perceptual Variables (Sleep Quality, Muscle Soreness, Stress Levels)
- Physical Performance Measures (Squat 1RM, 15 m × 17 Shuttle Run, 5.8 m × 6 Shuttle Run)
- Physiological Status Indicators (Urine Protein, Urobilinogen, Urine pH, Urine Specific Gravity)
-Data imputation
-Data normalized (Z-score)
NR - SMOTE
- 10-fold cross-validation
Lu et al. (2022) Any Muscle (traumatic) - Demographics and Injury History (Recent hamstring injury, Recent back injury, Age)
- Game Performance Metrics (Field goal percentage, 3-point shots made per game, 3-point shots attempted per game, Usage percentage, Offensive win share, Defensive win share)
Not Reported RFE (Recursive Feature Elimination) 10-fold cross-validation
Huang et al. (2023) Any region (overuse) - Psychological and Perceptual Variables (RPE: Ratings of Perceived Exertion)
- Physical Performance Measures (Double under, Squat, Bench press, Shuttle run, Sprint)
- Physiological Status Indicators (Instantaneous Heart Rate, Heart Rate Recovery, Protein, pH)
-Data imputation
-Data normalized (Z-score)
LDA (Linear Discriminant Analysis) - SMOTE
- 5-fold cross-validation
Javier Robles-Palazon et al. (2023) Soft tissue (traumatic) - Demographics and Injury History (Age, BMI, Injury history)
- Psychological and Perceptual Variables (Anxiety, Motivation, Team cohesion)
- Physical Performance Measures (Joint Range of Motion (ROM), Balance and Stability, Functional Performance Measures)
-Data imputation
-Weka software
Attribute Selected Classifier - Under-sampling Bagging
- 5-fold cross-validation
Kolodziej et al. (2023) Any region (traumatic) - Demographics (Age, Height, Weight)
- Physical Performance Measures (Postural Control and Balance, Strength Measures, Joint Kinematics, Joint Moments)
- Ground Reaction Forces (Peak vGRF)
Data normalized LASSO (Least Absolute Shrinkage and Selection Operator) 15-fold cross-validation
NR, not reported; HSI, hamstring strain injuries.