Classifier for notifications of environmental accidents in the National System of Environmental Emergencies of IBAMA

• Filipe de M. Santos UFPB
• Lucas M. Aguiar UFPB
• Gilberto F. de Sousa Filho UFPB
• Bruno J. Sousa Pessoa UFPB

DOI: https://doi.org/10.5753/wcge.2023.230024

Abstract

Siema is the IBAMA system responsible for registering reports of environmental accidents from citizens throughout Brazil. According to IBAMA’s own database, 28% of the reports are wrong, generating a financial and human cost for the institution during the verification of these supposed occurrences. This work proposes the use of machine learning techniques to classify future reports of accidents as valid or not, aiming at the best use ofpublic resources in the analysis of notifications received by the agency. Three machine learning models were employed and classification metrics were presented regarding Siema records, obtaining a classifier that is able to correctly identify 91% of invalid accident reports, in addition to an overall accuracy of 89%.

Keywords: Environmental accidents, Machine learning, Interpretability

References

Abu-Mostafa, Y. S., Magdon-Ismail, M., and Lin, H.-T. (2012). Learning From Data. AMLBook.

Breiman, L., Friedman, J., Stone, C., and Olshen, R. (1984). Classification and Regression Trees. Taylor & Francis.

Chand, S. and Zhang, Y. (2022). Learning from machines to close the gap between funding and expenditure in the australian national disability insurance scheme. International Journal of Information Management Data Insights, 2(1):100077.

Cichosz, P. (2011). Assessing the quality of classification models: Performance measures and evaluation procedures. Open Engineering, 1(2):132–158.

du Boisberranger, J., Van den Bossche, J., Estève, L., and J. Fan, T. (2022). scikit-learn: machine learning in python.

Frempong, N. K., Nicholas, N., and Boateng, M. (2017). Decision tree as a predictive modeling tool for auto insurance claims. International Journal of Statistics and Applications, 2017:117–120.

Haritsah Luthfi, N. and Hartoyo, A. (2023). Ai explanation related covid hoax detection using support vector machine and logistics regression methods. Jurnal Media Informatika Budidarma, 7:170–177.

Severino, M. K. and Peng, Y. (2021). Machine learning algorithms for fraud prediction in property insurance: Empirical evidence using real-world microdata. Machine Learning with Applications, 5:100074.

YÜCEL, A. (2022). A novel data processing approach to detect fraudulent insurance claims for physical damage to cars. Journal of New Results in Science, 11:120–131.