This repository contains the R code for the analysis performed in the paper "Forecasting Brazilian mortality rates due to occupational accidents using autoregressive moving average approaches"

Forecasting Brazilian mortality rates due to occupational accidents using autoregressive moving average approaches

DOI: https://doi.org/10.1016/j.ijforecast.2020.09.010

Abstract: We examine the mortality rates due to occupational accidents of the three states in the southern region of Brazil using the autoregressive integrated moving average (ARIMA), beta autoregressive moving average (ARMA), and Kumaraswamy autoregressive moving average (KARMA) models to fit the data sets, considering monthly observations from 2000 to 2017. We compare them to identify the best predictive model for the southern region of Brazil. We also provide descriptive analysis, revealing the victims vulnerability characteristics and comparing them between the states. A clear increase was seen in female participation in the labor market, but the number of deaths from occupational accidents did not increase by the same proportion. Moreover, the state of Paran stood out for having the highest mortality rate from work-related accidents. The fitted ARIMA and ARMA models using a 6-month time frame presented similar accuracy measurements, while KARMA performed the worst.

Keywords: Fatal work-related accidents; ARIMA; ARMA; KARMA; Forecasting; Time series

How to cite this work

Melchior, Cristiane; Zanini, Roselaine Ruviaro; Guerra, Renata Rojas; Rockenbach, Dinei A.; Forecasting Brazilian mortality rates due to occupational accidents using autoregressive moving average approaches, International Journal of Forecasting, vol 37, issue 2, 2021, pp 825-837, DOI: 10.1016/j.ijforecast.2020.09.010.

bibtex @article{MELCHIOR2021825, author = {Cristiane Melchior and Roselaine Ruviaro Zanini and Renata Rojas Guerra and Dinei A. Rockenbach}, title = {Forecasting Brazilian mortality rates due to occupational accidents using autoregressive moving average approaches}, journal = {International Journal of Forecasting}, volume = {37}, number = {2}, pages = {825-837}, year = {2021}, issn = {0169-2070}, doi = {https://doi.org/10.1016/j.ijforecast.2020.09.010}, url = {https://www.sciencedirect.com/science/article/pii/S0169207020301515}, keywords = {Fatal work-related accidents, ARIMA, ARMA, KARMA, Forecasting, Time series}, abstract = {We examine the mortality rates due to occupational accidents of the three states in the southern region of Brazil using the autoregressive integrated moving average (ARIMA), beta autoregressive moving average (ARMA), and Kumaraswamy autoregressive moving average (KARMA) models to fit the data sets, considering monthly observations from 2000 to 2017. We compare them to identify the best predictive model for the southern region of Brazil. We also provide descriptive analysis, revealing the victims vulnerability characteristics and comparing them between the states. A clear increase was seen in female participation in the labor market, but the number of deaths from occupational accidents did not increase by the same proportion. Moreover, the state of Paran stood out for having the highest mortality rate from work-related accidents. The fitted ARIMA and ARMA models using a 6-month time frame presented similar accuracy measurements, while KARMA performed the worst.} }