CHOICE OF MATHEMATICAL INSTRUMENT FOR MODEL OF FORECASTING OF UNFAVORABLE AIRCRAFT ACCIDENTS IN THE FLIGHT
DOI:
https://doi.org/10.26906/SUNZ.2018.5.020Keywords:
forecasting, classification, Bayes classifier, support vector machine, convolutional neural network, recurrent neural networkAbstract
The purpose of the article. Carrying out research and selection of the most effective mathematical device for constructing the model of forecasting of adverse aviation events during the flight. Results. The paper analyzes the known methods used to solve data classification problems. This is necessary to determine the feasibility of their use for building a model of forecasting unfavorable aircraft accidents in the flight, based on the analysis of text messages. The following methods are considered: logistic regression, support wind method, naive Baes classifier, random forest. In addition, to solve this class of problems, convolutional and recurrent neural networks using deep learning algorithms are considered. Conclusions. As a result of the analysis of these methods, a mathematical instrument of deep neural networks was chosen to build a model of forecasting of unfavorable aircraft accidents in the flight, based on the analysis of text messages. Due to the application of their deep learning algorithms, they have the highest accuracy compared to traditional approaches.Downloads
References
Состояние безопасности полетов в гражданской авиации государств-участников соглашения о гражданской авиации и об использовании воздушного пространства в первом полугодии 2018 г. [Электронный ресурс]. – Режим доступа: https://mak-iac.org/upload/iblock/5b0/bp-18-1.pdf.
П.Флах Машинное обучение. Наука и искусство построения алгоритмов, которые извлекают знания из даннях / пер. с англ А.А.Слинкина. ̶М.: ДМК Пресс, 2015. –̶400 с.
Бідюк П.І. Основні етапи побудови і приклади застосування мереж Байєса / П.І. Бідюк, Н.В. Кузнєцова // Систем. дослідж. та інформ. технології. — 2007. — № 4. — С. 26 — 39.
dos Santos C.N., Gatti M. Deep Convolutional Neural Networks for Sentiment Analysis of Short Texts // COLING, 2014. P. 69 – 75.
Y. Kim. Convolutional neural networks for sentence classification. arXiv:1408.5882 [cs.CL], 2014.
C. Olah. Neural networks, recurrent neural networks, convolutional neural networks [Электронный ресурс]. – Режим доступа: http://colah.github.io.
Hochreiter S., Schmidhuber J. Long short-term memory // Neural computation 9, 1997. Issue 8. P. 1735 – 1780.
