Spark-based multi-verse optimizer as wrapper features selection algorithm for phishing attack challenge

Nowadays, phishing attacks have grown rapidly, and there is an urgent need to introduce a suitable detection method that has the ability to detect different types of phishing attacks. This paper investigates the capability to use bio-inspired meta-heuristic algorithms to improve the performance of the detection engine for phishing attacks by reducing the number of features. This improvement was practiced by investigating the effectiveness of five meta-heuristic algorithms: Particle Swarm Optimization (PSO), Firefly Optimization Algorithm (FFA), Multi-Verse Optimizer (MVO), Moth-Flame Optimization algorithm (MFO), and BAT optimization algorithm, to select the relevant features that could be affected directly by different types of phishing attacks. The suggested detection model was tested and evaluated using four benchmark phishing attack datasets, and the Apache Spark-based decision tree algorithm was selected as a detection engine. The conducted experiments have demonstrated that the Spark-based MVO algorithm achieved the highest detection rate for detecting different types of phishing attacks within four phishing attack datasets. Moreover, the suggested detection model was able to reduce effectively the feature space, which could enhance the computational efficiency.