Analysis of road accident mortality based on time of occurrence for Kerala, India
DOI:
https://doi.org/10.18203/2394-6040.ijcmph20233791Keywords:
Road accident, Time series regression, Prophet, Deep learning, SeasonalityAbstract
Background: Road traffic accidents (RTA) pose a significant socio-economic burden and global public health concern. Monitoring road safety initiatives' efficacy necessitates analysing RTA incidence. This study examines time zone-specific RTA mortality in Kerala state, India, from 2016 to 2021.
Methods: Utilizing compiled secondary-level time series data, the study encompasses total RTA fatalities in Kerala from 2016 to 2021. Data includes fatalities per year in nine consecutive three-hour time periods. Exploratory data analysis, time series regression, and exponential smoothing were employed for analysis.
Results: Data reveals fluctuating trends in road accident (RA) fatalities, peaking in 2018 with a notable decrease in 2020. 18:00 to 21:00 recorded the highest and lowest fatalities, total 901 deaths. Disproportionate RA fatalities occurred from 06:00 to 09:00 (527 deaths) and 15:00 to 18:00 (697.5 deaths). The study employs Holt-Winters exponential smoothing for short-term forecasting, with a mean absolute scaled error (MASE) less than 1 signifying accurate predictions.
Conclusions: The analysis highlights temporal patterns, emphasizing 18:00 to 21:00 as critical. Holt-Winters exponential smoothing proves vital for accurate short-term forecasting, with MASE reflecting precision. Urgency is stressed in adopting targeted measures for time-specific road accidents. Government intervention is pivotal, advocating for improved infrastructure, enhanced driver education, efficient vehicle management, and sustained traffic enforcement. Tailoring traffic laws to time zones, coupled with forecasting techniques, aligns with the overarching goal of enhancing road safety and reducing RA mortality rates.
Metrics
References
Nazari SS, Moradi A, Rahmani K. A systematic review of the effect of various interventions on reducing fatigue and sleepiness while driving. Chinese J traumatol. 2017;20(5):249-58.
Sunny CM, Nithya S, Sinshi KS, Vinodini V, Kg AL, Anjana S, Manojkumar TK. Forecasting of road accident in Kerala: A case study. In: 2018 International Conference on Data Science and Engineering (ICDSE). IEEE. 2018;1-5.
Jayan KD, Ganeshkumar B. Identification of accident hot spots: AGIS based implementation for Kannur District, Kerala. Int J Geomat Geosci. 2010;1(1):51-9.
Mohan D. Road accidents in India. IATSS Research. 2009;33(1):75.
Behura A, Behura A. Road accident prediction and feature analysis by using deep learning. In2020 11th International Conference on Computing, Communication and Networking Technologies (ICCCNT). IEEE. 2020;1-7.
Sumit K, Ross V, Ruiter RA, Polders E, Wets G, Brijs K. An exploration of characteristics and time series forecast of fatal road crashes in Manipal, India. Sustainability. 2022;14(5):2851.
Abdulhafedh A. Road crash prediction models: different statistical modeling approaches. J Transport Technol. 2017;7(02):190.
Vipin N, Rahul T. Road traffic accident mortality analysis based on time of occurrence: Evidence from Kerala, India. Clin Epidemiol Global Health. 2021;11:100745.
Deretić N, Stanimirović D, Awadh MA, Vujanović N, Djukić A. SARIMA modelling approach for forecasting of traffic accidents. Sustainability. 2022;14(8):4403.
Milton JC, Shankar VN, Mannering FL. Highway accident severities and the mixed logit model: an exploratory empirical analysis. Accident Analysis & Prevention. 2008;40(1):260-6.