Exploiting Machine Learning Models to Avoid Texting While Driving

Author(s): Torres, Renato H.; Ohashiz, Orlando; Garcia, Gabriel; Rocha, Filipe; Azpúrua, Héctor; Pessin, Gustavo
Summary: Text while driving is a worldwide phenomenon which is acknowledged as a meaningful problem for road safety. Text while driving is one of the most dangerous types of
distraction because it involves visual, cognitive and physical distraction. Due to the advancement of technology, in the same way that it increases the possibilities of drivers’ distractions, also increases the number of proposals that aim to avoid this behavior. In this context, we propose an intelligent system to identify text while driving. We deploy four machine learning models: Decision Tree; Support Vector Machine; Random Forest and Gradient Boosting. We compare their performances and significance for the domain of this work. The evaluation of the models were carried out with a real word collected dataset. Regarding the accuracy of the models, Random Forest and Gradient Boosting presented performance superior to 0:93. Recall ( 0:94) and Precision ( 0:94) also presented good results. In this work, these two metrics are highlighted because they are related, respectively, to driver’s safety and passenger’s convenience. In addition, we also carried out an analysis of the social and economic impacts that the proposed model can cause.
Journal: International Joint Conference on Neural Networks - IJCNN 2019
Year: 2019
Pages: p. 1-8
DOI: https://doi.org/10.1109 / IJCNN.2019.8852202
Year of publication: 2019
Available at: https://www.researchgate.net/publication/336165494_Exploiting_Machine_Learning_Models_to_Avoid_Texting_While_Driving
Publisher and ISSN: IEEE