Date of Award
Senior Honors Thesis
College or School
Program or Major
Data Science & Analytics
Bachelor of Science
Machine learning models can be trained to classify time series based sports motion data, without reliance on assumptions about the capabilities of the users or sensors. This can be applied to predict the count of occurrences of an event in a time period. The experiment for this research uses lacrosse data, collected in partnership with SPAITR - a UNH undergraduate startup developing motion tracking devices for lacrosse. Decision Tree and Support Vector Machine (SVM) models are trained and perform with high success rates. These models improve upon previous work in human motion event detection and can be used a reference in future work where similar models are developed to detect and count events in other sports data, or human motion data in general.
Cashman, Mallory, "Realtime Event Detection in Sports Sensor Data with Machine Learning" (2022). Honors Theses and Capstones. 699.
Categorical Data Analysis Commons, Data Science Commons, Graphics and Human Computer Interfaces Commons, Longitudinal Data Analysis and Time Series Commons, Other Computer Sciences Commons, Other Mathematics Commons, Other Statistics and Probability Commons, Sports Management Commons, Sports Sciences Commons, Statistical Models Commons, Technology and Innovation Commons