Body Performance Analysis with Machine Learning and ANOVA Methods

Detection of body performance with artificial intelligence using body data can provide a precise and objective evaluation of individuals' physical abilities. Artificial intelligence models can create personalized training programs and development plans by learning from large data sets. This can help athletes and fitness enthusiasts optimize their performance and reduce their risk of injury. Additionally, health professionals and coaches can make more effective and targeted interventions by making data-based decisions. In this study, it was aimed to predict the performances of athletes using body data. A dataset containing 13,393 rows of data in total was used. There are four classes in the dataset and they represent performance levels. Artificial Neural Network (ANN), Gradient Boosting (GB), Random Forest (RF) machine learning methods were used to classify the data. The cross validation method was used to objectively evaluate the results of the models. Confusion matrix and performance metrics were used to analyze the performance of the models. Classification successes and other performance metrics obtained as a result of the classification of the models were used to compare the performances of the models. The highest classification success of 74.5% was obtained from the ANN model. The lowest classification success was obtained from the RF model with 69.6%. ANOVA was used to examine the effects of the features in the dataset used on classification. The effects of the features were analyzed and their importance level was determined. It is thought that the proposed models can be used in applications by using them in performance analysis.