Machine Learning of Pacing Patterns for Half Marathon

• Every year over 40 000 runners participate in Gothenburg Half Marathon, one of the world’s largest half-marathons. Based on publicly available results data (423 496 entries) for ten years (2010 – 2019), we investigate machine learning models for two tasks: prediction of finishing times and identification of runners risking hitting the wall. Our models improve results over the current baseline on finish time prediction and manage to correctly identify many of the runners who later hit the wall, although it also misclassifies many who do not.

Subject headings

NATURVETENSKAP  -- Data- och informationsvetenskap -- Annan data- och informationsvetenskap (hsv//swe)

NATURAL SCIENCES  -- Computer and Information Sciences -- Other Computer and Information Science (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Datorsystem (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Computer Systems (hsv//eng)

Keyword

Machine Learning

Half Marathon

Pacing Pattern

Performance Analysis

Publication and Content Type

art (subject category)

vet (subject category)