Classification of DoS Attacks in IoT using Different Feature Selection Methods and Deep Learning

Firas Raad Sultan; Islam R. Abdelmaksoud; Hazem M. El-Bakry

This research paper presents a method for classifying Denial of Service (DoS) attacks. The approach utilizes the XGBoost algorithm to select the features and assesses the performance of two classification models: Naive Bayes and Recurrent Neural Network (RNN). The main goal is to improve the accuracy and efficiency of DoS attack classification. By employing the XGBoost algorithm we can identify the features for accurate classification. To evaluate the performance of Naive Bayes and RNN models a comprehensive dataset of DoS attacks is considered, encompassing attack types and their corresponding features. Prior to analysis the dataset undergoes preprocessing measures to ensure data quality. Experimental results provide evidence supporting the effectiveness of our proposed approach. The utilization of XGBoost algorithm significantly enhances classification performance, for both Naïve Bayes and RNN models. The RNN model, with its ability to capture temporal patterns, outperforms Naive Bayes. This research contributes to the field by offering an innovative approach that combines feature selection with two distinct classification models. The proposed method provides a valuable framework for accurately identifying and categorizing DoS attacks, aiding network administrators and security professionals in mitigating the impact of such attacks.

Select Volume / Issues:

• Volume 15, Issue 1, Jan. 2024

Year:

2024

Type of Publication:

Article

Keywords:

DoS Attack; XGboost; Data Mining; Network Security; Recurrent Neural Network; Naive Bayes; Intrusion Detection System

Journal:

IJECCE

Volume:

15

Number:

1

Pages:

1-17

Month:

January

ISSN:

2249-071X