A Predictive Classification Model For Running Injury
Loading...
Date
2022-07-25
Authors
Ganesan, Devesh Raj
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
Running- related injury is musculoskeletal pain in the lower limbs that causes a
restriction on or stoppage of running. Running injuries have been collectively studied
in terms of the attributing factors as well as faulty trainings. Various models have been
devised to address this issue, however the percentage of running injury occasions are
still alarming. Studies have yet to develop a good predictive classification model for
running injury. Therefore, the goal of this study was to identify the determinants of
running injuries, to classify running data by degree of severity and to develop a
predictive classification model of running injury. Two case studies related to running
injury were retrieved from the public available domain. Data mining approach was
conducted to pre-process and to classify data into three injury levels: low, moderate,
and severe risks aided by Waikato Environment for Knowledge Analysis (WEKA)
version 3.8.6 tool. The J48, SMO, Random Forest, and Simple Logistic algorithms were
used for 10-fold cross validation mode classification benchmarked on the ZeroR
baseline algorithm. Findings reveal that classification accuracy obtained were from
70% to 100%.