Seybold Report ISSN: 1533-9211
T.Elangovan
Ph.D. Research Scholar, Dept. of Computer Science Erode Arts and Science College,
Erode-638 009, Tamilnadu, India. E-mail: elangovaneasc@gmail.com
Dr.V.Sheshathri
Assistant Professor, Department of Computer Applications, Erode Arts and Science College,
Erode-638 009, Tamilnadu, India. E-mail: sheshathrieac@gmail.com
Dr.S.Sukumaran
Associate Professor in Computer Science, Erode Arts and Science College,
Erode-638 009, Tamilnadu, India. E-mail: prof_sukumar@yahoo.co.in
Vol 17, No 10 ( 2022 ) | Licensing: CC 4.0 | Pg no:2165-2176 | Published on: 31-10-2022
Abstract
Objectives: Cyber security is the provision of an effective cyber threats detection technique based on deep neural networks. Cyber attacks is highly challenging to protect the systems against threats and malicious behaviors in networks. Methods/Statistical Analysis: The proposed technique input data first preprocessed, that initial data format convert into image format. Then data normalization process eliminates the different dimensional data. Then mean convolution layer and convolutional neural network based layer to find the cyber threats from normal data. This method recommends a new Enhanced Convolutional Neural Network architecture type known as Mean Convolution Layer (ECNN-MCL) developed for learning the anomalies content features and then identifying the particular abnormality. Findings: The recommended ECNN-MCL helps in designing a strong network intrusion detection system that includes an innovative form of convolutional layer that can teach low level abnormal characteristics. The system emphases on discriminating between true positive and false positive alerts, thus helping security analysts to fast respond to cyber threats. Applications/Improvements: The experiment in this method is performed using NSLKDD data set. The performance comparison is compared with proposed ECNN-MCL and existing methods of SVM, k-NN, and CNN. The experimental results ensure that our proposed method is capable of being employed as learning-based model for network intrusion-detection. Finally, the experimentation results are shown through Mat Lab R2013a.
Keywords:
Intrusion Detection, Deep Neural Networks and Cyber Threat Detection, ECNN-MCL, Image Format, Normalization