Use of Machine Learning to Model Volume Load Effects on Changes in Jump Performance

Area of Science:

• Sports Science
• Biomechanics
• Artificial Intelligence in Sports

Background:

• Understanding the relationship between training load and athletic performance is crucial for optimizing training programs.
• Countermovement jump (CMJ) is a key performance indicator in many sports, particularly track-and-field.
• Predicting changes in CMJ performance based on training volume load (VL) can aid in athlete development.

Purpose of the Study:

• To develop and utilize an artificial neural network (ANN) to model the effects of 15 weeks of resistance training on CMJ performance.
• To analyze the association between weekly resistance training volume load (VL) and changes in CMJ height in male track-and-field athletes.

Main Methods:

• Collected 15-week resistance training volume load (VL) data from 21 NCAA Division I male track-and-field athletes.
• Measured weekly countermovement jump (CMJ) height to determine overall performance changes.
• Employed a feed-forward artificial neural network (ANN) with 5 hidden layers to model the relationship between weekly VL and CMJ height changes.

Main Results:

• The developed ANN accurately predicted individual changes in CMJ height, with an average error ranging from 0.21 to 1.47 cm.
• The model demonstrated that 15 weeks of VL data could effectively capture the variability in CMJ performance changes.
• Analysis revealed that volume loads during deload or taper weeks were the most significant predictors (10%-17%) of changes in CMJ performance.

Conclusions:

• Artificial neural networks (ANNs) provide an effective method for modeling the impact of weekly training volume load (VL) on countermovement jump (CMJ) performance.
• ANNs can be utilized to identify the relative importance of specific training periods, such as deload and taper weeks, in predicting performance outcomes.
• This approach offers valuable insights for periodization and training load management in track-and-field athletes.