Artificial Hummingbird Algorithm with Deep Variational Autoencoder Driven Intrusion Detection in Big Data Analytics Environment

Abstract

Recently, the enormous quantities of data and its linear upsurge have modified the importance of data safety and data analysis techniques for Big Data (BD). Intrusion Detection System (IDS) is a method that examining and monitoring data to identify some intrusion within the network or system. Higher volume, variability, and maximal speed of data produced from the network that can be achieved in the data analysis method to identify attacks by conventional methods lead to more complexity. BD methods have been employed in IDS to manage BD for efficient and accurate procedure of data analysis. Deep Learning (DL) and BD approaches are incorporated to increase the performance of IDSs. This manuscript offers an Artificial Hummingbird Algorithm with Artificial Intelligence Driven Intrusion Detection in Big Data Environment (AHAAI-IDBDE) technique. The purpose of the AHAAI-IDBDE technique exploit the Feature Selection (FS) with hyperparameter selection strategy for intrusion detection. For handling BD, MapReduce is used. In the AHAAI-IDBDE technique, Z-score normalization has been used to scale the input data. For electing an optimal set of features, the Binary Volleyball Premier League (BVPL) can be used. To detect intrusions, the AHAAI-IDBDE technique uses Deep Variational Autoencoder (DVAE) model. For hyperparameter selection process, the Artificial Hummingbird Algorithm (AHA) can be applied. The experimental value of the AHAAI-IDBDE method has been examined deploying benchmark IDS dataset. The extensive comparison analysis portrayed the betterment of the AHAAI-IDBDE technique with other systems.