Providing Explainable Race-Time Predictions and Training Plan Recommendations to Marathon Runners

Abstract

Millions of people participate in marathon events every year, typically devoting at least 12-16 weeks to building their endurance and fitness so that they can safely complete these gruelling 42.2km races. Most runners follow a training plan that is tailored to their expected finish-time (e.g. sub-4 hours or 4-5 hours), and these plans will prescribe a complex mixture of training sessions to help them achieve these times. However, such plans cannot adapt to the individual needs (fitness levels, changing goals, personal preferences) of runners, providing only broad training guidance rather than more personalised support. The development of wearable sensors and mobile fitness applications facilitates the collection of a large amount of training data from runners. In this paper, we propose a recommender system that utilizes such training data to deliver more personalised training advice to runners, using ideas from casebased reasoning to reuse and adapt the training habits of similar runners. Explainability plays a significant role in this type of system, and we also describe how the predictions and recommendation advice can be presented to runners. An initial off-line evaluation is presented based on a large-scale, real-world dataset. CCS CONCEPTS • Computing methodologies → Machine learning algorithms; Feature selection; • Applied computing → Health informatics.

How recreational marathon runners hit the wall: A large-scale data analysis of late-race pacing collapse in the marathon

“…Although such data was not available in our dataset, the increasingly widespread adoption of mobile devices, smartwatches, and wearable sensors [55,56] has the capacity to generate large volumes of additional data (heart-rate, cadence, and power), which may be useful in this regard in the future [57,58]. Already, the availability of such diverse sources of data is enabling several new types of health and fitness applications [59][60][61][62][63] and the emergence of powerful new machine learning techniques has been used to support a variety of related prediction and planning tasks in several sporting domains [64][65][66][67][68][69][70][71][72][73] It is also worth noting that the model of the wall analysed here is defined by a pair of parameters-degree of slowdown and length of slowdown-with specific values-0.25 and 5km, respectively-and it is reasonable to question whether the results would be different if different values had been chosen. We have considered several alternative sets of values and, within reasonable levels of tolerance, there is no material change to the nature of the results as presented.…”

Sports Recommender Systems: Overview and Research Issues

“…Case-based recommendation (CBR) can be applied, for example, to infer training practices of persons with a similar running performance. Approaches to the CBR of training sessions for marathon runners are presented in Berndsen et al [15] (training plans) and Feely et al [46,47] (training plans and race time predictions). Follow-up works focus on the "inclusion" of recreational marathon runners, for example, in terms of realistic and reasonable training plans [49,113].…”

An Overview of Recommendation Techniques and Their Applications in Healthcare

“…Interpretability: The most common view of interpretability in RS is to increase the transparency of algorithms [14], [15], [40], [43], [164], which is especially important in health RS. Reliable explanations can greatly improve end-users' confidence in the recommendation results [126].…”