Simplifying External Load Data in NCAA Division-I Men's Basketball Competitions: A Principal Component Analysis

Area of Science:

• Sports Science
• Biomechanics
• Data Analysis

Background:

• Understanding external load in NCAA Division I basketball is crucial for performance and injury prevention.
• Existing methods for analyzing athlete load data can be complex and require simplification.

Purpose of the Study:

• To simplify external load data from NCAA Division I basketball games using Principal Component Analysis (PCA).
• To assess if PCA-derived load metrics can differentiate between player positions (POS).

Main Methods:

• Collected external load data from 10 NCAA Division I men's basketball athletes using inertial measurement units.
• Applied Principal Component Analysis (PCA) to simplify a large dataset of external load variables.
• Utilized multinomial logistic regression to predict player positions based on PCA-identified load metrics.

Main Results:

• PCA identified two factors explaining 81.42% of the total variance, simplifying the external load data.
• Maximum speed (maxSPD), total decelerations (totDEC), total jump load (totJUMP), and total mechanical load (totMECH) were significant predictors of player position.
• The model showed high accuracy in predicting positions (AUC > 0.80), though differentiating between guards and forwards presented challenges due to overlapping demands.

Conclusions:

• PCA effectively simplifies complex external load data in NCAA Division I basketball.
• Key load variables (maxSPD, totDEC, totJUMP, totMECH) are sensitive to positional differences.
• These simplified metrics can inform individualized training and recovery strategies for basketball athletes.