Recommendations for marathon runners: on the application of recommender systems and machine learning to support recreational marathon runners.

Every year millions of people, from all walks of life, spend months training to run a traditional marathon. For some it is about becoming fit enough to complete the gruelling 26.2 mile (42.2 km) distance. For others, it is about improving their fitness, to achieve a new personal-best finish-time. In this paper, we argue that the complexities of training for a marathon, combined with the availability of real-time activity data, provide a unique and worthwhile opportunity for machine learning and for recommender systems techniques to support runners as they train, race, and recover. We present a number of case studies-a mix of original research plus some recent results-to highlight what can be achieved using the type of activity data that is routinely collected by the current generation of mobile fitness apps, smart watches, and wearable sensors.

Prediction Equations for Marathon Performance: A Systematic Review.

PURPOSE: Despite the volume of available literature focusing on marathon running and the prediction of performance, no single prediction equations exists that is accurate for all runners of varying experiences and abilities. Indeed the relative merits and utility of the existing equations remain unclear. Thus, the aim of this study was to collate, characterize, compare, and contrast all available marathon prediction equations. METHODS: A systematic review was conducted to identify observational research studies outlining any kind of prediction algorithm for marathon performance. RESULTS: Thirty-six studies with 114 equations were identified. Sixty-one equations were based on training and anthropometric variables, whereas 53 equations included variables that required laboratory tests and equipment. The accuracy of these equations was denoted via a variety of metrics; r2 values were provided for 68 equations (r2 = .10-.99), and an SEE was provided for 19 equations (SEE 0.27-27.4 min). CONCLUSION: Heterogeneity of the data precludes the identification of a single "best" equation. Important variables such as course gradient, sex, and expected weather conditions were often not included, and some widely used equations did not report the r2 value. Runners should therefore be wary of relying on a single equation to predict their performance.