Detecção de anomalias em poços produtores de petróleo usando aprendizado de máquina

2020 ◽  
Author(s):  
Wander Fernandes Júnior ◽  
Ricardo Emanuel Vaz Vargas ◽  
Karin Satie Komati ◽  
Kelly Assis de Souza Gazolli

Anomalias em poços produtores de petróleo podem provocar impactos financeiros significativos. O uso de aprendizado de máquina para detectar essas situações podem prevenir interrupções indesejadas de produção bem como custos de manutenção. Nesse contexto, este trabalho propõe a aplicação e comparação de classificadores para detecção de anomalias em poços de produção de petróleo e gás. Classificadores de classe única Floresta de Isolamento, \textit{One-class Support Vector Machine} (OCSVM), \textit{Local Outlier Factor} (LOF) e Envelope Elíptico foram aplicados em uma base de dados com casos reais, sendo o melhor desempenho obtido pelo LOF com medida F1 de 88,2\%, seguido da Floresta de Isolamento com 74,3\%. Os resultados obtidos apresentam melhoria em comparação ao \textit{benchmark} de referência e estimulam a continuação do trabalho com a experimentação de outras famílias de classificadores.

Sensors ◽  
2019 ◽  
Vol 19 (14) ◽  
pp. 3188 ◽  
Author(s):  
Vitor Hugo Bezerra ◽  
Victor Guilherme Turrisi da Costa ◽  
Sylvio Barbon Junior ◽  
Rodrigo Sanches Miani ◽  
Bruno Bogaz Zarpelão

Internet of Things (IoT) devices have become increasingly widespread. Despite their potential of improving multiple application domains, these devices have poor security, which can be explored by attackers to build large-scale botnets. In this work, we propose a host-based approach to detect botnets in IoT devices, named IoTDS (Internet of Things Detection System). It relies on one-class classifiers, which model only the legitimate device behaviour for further detection of deviations, avoiding the manual labelling process. The proposed solution is underpinned by a novel agent-manager architecture based on HTTPS, which prevents the IoT device from being overloaded by the training activities. To analyse the device’s behaviour, the approach extracts features from the device’s CPU utilisation and temperature, memory consumption, and number of running tasks, meaning that it does not make use of network traffic data. To test our approach, we used an experimental IoT setup containing a device compromised by bot malware. Multiple scenarios were made, including three different IoT device profiles and seven botnets. Four one-class algorithms (Elliptic Envelope, Isolation Forest, Local Outlier Factor, and One-class Support Vector Machine) were evaluated. The results show the proposed system has a good predictive performance for different botnets, achieving a mean F1-score of 94% for the best performing algorithm, the Local Outlier Factor. The system also presented a low impact on the device’s energy consumption, and CPU and memory utilisation.


2020 ◽  
Author(s):  
V Vasilevska ◽  
K Schlaaf ◽  
H Dobrowolny ◽  
G Meyer-Lotz ◽  
HG Bernstein ◽  
...  

2019 ◽  
Vol 15 (2) ◽  
pp. 275-280
Author(s):  
Agus Setiyono ◽  
Hilman F Pardede

It is now common for a cellphone to receive spam messages. Great number of received messages making it difficult for human to classify those messages to Spam or no Spam.  One way to overcome this problem is to use Data Mining for automatic classifications. In this paper, we investigate various data mining techniques, named Support Vector Machine, Multinomial Naïve Bayes and Decision Tree for automatic spam detection. Our experimental results show that Support Vector Machine algorithm is the best algorithm over three evaluated algorithms. Support Vector Machine achieves 98.33%, while Multinomial Naïve Bayes achieves 98.13% and Decision Tree is at 97.10 % accuracy.


2011 ◽  
Vol 131 (8) ◽  
pp. 1495-1501
Author(s):  
Dongshik Kang ◽  
Masaki Higa ◽  
Hayao Miyagi ◽  
Ikugo Mitsui ◽  
Masanobu Fujita ◽  
...  

Author(s):  
Ryoichi ISAWA ◽  
Tao BAN ◽  
Shanqing GUO ◽  
Daisuke INOUE ◽  
Koji NAKAO

Sign in / Sign up

Export Citation Format

Share Document