Statistical and Transfer Learning Model to Analyze Endurance Performance with Aging

The decline of human physical ability with aging is an important factor to consider as to athletic performance evaluation. Physiological factors or real-world performance data have been used to build mathematical models. Marathon is a sports event attracting more and more participants. However, due to the limitation of field size, the race organizers usually must qualify each athlete for certain events based on age and gender. We use statistical and machine learning models for analyzing the Boston Marathon results to delineate the decrease in elder athletes' endurance performance. We also apply the transfer learning model on the mobile activity tracking social media platform Strava to classify beginners or professional athletes using performance data and applying sentiment classification to athletes' comments. Our results may provide valuable insights for athletes, coaches, and governing bodies for athletics.