Prediction of physical fitness and performance of Wushus athletes based on machine learning and fuzzy TOPSIS method

Predicting the fitness of athletes is important for improving their performance over time. Training sessions and past performance records are the common features used to guide athletes’ gradual improvement. This study integrates conventional deep learning and partial fuzzy TOPSIS to assess athletes’ physical fitness and performance. First, the learning process identifies the precise demands needed for ongoing improvement through optimized training. The model regularly checks whether the training inputs meet the evolving needs of different sessions. These identified inputs are further validated using the fuzzy TOPSIS method to determine a clear pathway for sustained, reliable performance gains. The prioritized results produced over multiple iterations are useful in identifying highly effective fitness programs that support steady improvement. The proposed method is evaluated using the Wushu training requirements dataset to identify specific training needs and performance outcomes based on multiple practice sessions.