Plantar Pressure Topography Based Sports Branch Prediction System Modeling in 7-11 Years Old Athletes

Abstract

Objective: The sole of foot plays a crucial role in sports movements, as it applies pressure to the ground and transfers loads. The foot pressing types vary depending on the sport played by the athlete. The aim of this study is to develop a model that can predict sports branches from the plantar pressure types of athletes. Methods: A total of 80 athletes, 54 athletics and 26 combat athletes, between the ages of 7-11 were included in the study where static pedobarographic measurements of the participants were collected. First we applied conventional statistical analysis on the featured obtained from the measurements of the data using Fisher Freeman Halton Exact test. Then, we implemented sports branch prediction based on the data obtained from these measurements using advanced machine learning and deep learning techniques. Results: There was no statistically significant difference between the plantar compression types of the participants according to the branches (p > .05). In the machine learning classification based on foot plantar compression, the best success was found to be 56.9% with Linear Support Vector Machine. When the branch prediction successes made with deep learning were examined, it was found that the average branch prediction was 82.58±7.62% in the foot with pes planus, 87.84±17.56% in the normal foot, and 85.95±21.19% in the foot with pes cavus. Conclusion: In the study, it was determined that the success of branch prediction made with machine learning techniques was low, and the success of deep learning was high. With the development of the method used in this study in future studies, an idea can be obtained about which branch of the foot plantar pressure type is more prone to and innovations can be brought to the branch selection methods.