Incident Detection in Industrial Processes Utilizing Machine Learning Techniques
This work provides a comparative analysis of the most popular and widely used classification algorithms applied in industrial processes, in order to tackle the issue of incident detection of abnormal situations. The proposed analysis is based on actual datasets from derived by the operation of a chemical process system situated at the premises of CERTH/CPERI. The evaluation of the tested methods is based on cross-validation using a series of Monte-Carlo simulations among several free parameters of tested classifiers and finally the application of the Adaptive Boosting technique. The experimental results are highly reliable as the accuracy reaches 98% and the F-measure metric achieves a score over 97%. Therefore, the detection of potential malfunctions is achieved using the proposed machine learning techniques.
KeywordsMachine learning techniques Incident detection AdaBoost
This work has been partially supported by the European Commission through the project HORIZON 2020-INNOVATION ACTIONS (IA)-636302-SATISFACTORY.
- 1.Boser, B.E., Guyon, I.M., Vapnik, V.N.: A training algorithm for optimal margin classifiers. In: 5th Annual Workshop on Computational Learning Theory, ACM, pp. 144–152 (1992)Google Scholar
- 2.Breiman, L.: Random forests – random features. Technical Report 567, Statistics Department, University of California, Berkeley, ftp://stat.berkeley.edu/pub/users/breiman
- 3.Freund, Y., Schapire, R.E.: A desicion-theoretic generalization of on-line learning and an application to boosting. In: Computational Learning Theory, Springer, pp. 23–37 (1995)Google Scholar
- 5.Ghouse, F., Natarajan, P., Subramanian, S.: Fault diagnosis of batch reactor using machine learning methods. Model. Simul. Eng. 2014 (2014), Article ID 426402Google Scholar
- 6.Gowthami, R., Vijayachitra, S.: Fault detection and diagnosis in continuous stirred tank reactor (CSTR). Int. J. Tech. Res. Appl. 3(2), 07–11 (2015)Google Scholar
- 10.Nath, S.V., Behara, R.S.: Customer churn analysis in the wireless industry: a data mining approach. In: Annual Meeting of the Decision Sciences Institute, pp. 505–510 (2003)Google Scholar
- 11.Papadopoulos, S., Drosou, A., Dimitriou, N., Abdelrahman, O.H., Gorbil, G., Tzovaras, D.: A BRPCA based approach for anomaly detection in mobile networks. In: 30th International Symposium on Computer and Information Sciences, vol. 363, pp. 115–124 (2015)Google Scholar
- 16.Thiprungsri, S., Vasarhelyi, M.A.: Cluster analysis for anomaly detection in accounting data: an audit approach. Int. J. Digit. Acc. Res. 11, 69–84 (2011)Google Scholar
- 17.Vafeiadis, T., Ioannidis, D., Krinidis, S., Ziogou, C., Voutetakis, S., Tzovaras, D., Likothanassis, S.: Real-time incident detection: An approach for two interdependent time series. In: 2016 24th European Signal Processing Conference (EUSIPCO), pp. 1418–1422 (2016)Google Scholar