Article Preview
TopIntroduction
Recommender systems are defined as decision support systems which enable users to select an item, product or service in different domains such as movies, conferences, etc. Consequently, recommender systems provide assistance to users so that they overcome the information overload/big data problem (Bobadilla, Ortega, Hernando, & Gutiérrez, 2013). Various types of traditional recommender systems such a Collaborative Filtering (CF), Content-Based Filtering (CBF), Context-Aware Recommender Systems (CARS), and Hybrid Recommenders have been widely discussed and elaborated in literature (Bobadilla et al., 2013) Additionally, research in recommender systems has witnessed an improvement of traditional recommender systems through the introduction of social recommender systems such as (Brusilovsky, Oh, López, Parra, & Jeng, 2017; Xia, Asabere, Liu, Chen, & Wang, 2017; Asabere, Acakpovi, & Michael, 2018; Asabere, Xia, Meng, Li, & Liu, 2015; Xia, Asabere, Liu, Deonauth, & Li, 2014; Asabere, Xia, Wang, Rodrigues, Basso, & Ma, 2014; Xia, Asabere, Rodrigues, Basso, Deonauth, & Wang, 2013; Asabere, Xu, Acakpovi, & Deonauth, 2021).
Internationally, the functionality and appropriateness of information systems infrastructure substantiate contemporary information society. Nevertheless, the proliferation of the Internet has introduced the “Information Overload” and “Big Data” syndromes, which have also consequently increased the number of cyber-attacks over the years due to the high 4Vs of Big Data, namely: volume, veracity, variety and velocity (Chen, Mao, & Liu, 2014). Although recommender systems have been used for product or service recommendation, it is worth exploiting the possibility of applying recommender systems in the area of cyber security due to the current global increase in cyber-attacks (Polatidis, Pimenidis, Pavlidis, & Mouratidis, 2017; Lyons, 2014; Kott, 2014; Ramaki & Atani, 2016). Normally, cyber-attackers exploit susceptibilities within a network and form attack paths from one asset to another until they have reached the asset they wish to harm (Polatidis et al., 2017; Lyons, 2014; Kott, 2014; Ramaki & Atani, 2016).
Traditionally, cyber security professionals have to wait for an attack to occur and then identify the attack. Consequently, cyber security professionals are disadvantaged in a cyber-attack situation due to the fact that they have to assiduously maneuver such attacks before the network is compromised (Polatidis et al., 2017; Lyons, 2014; Kott, 2014; Ramaki & Atani, 2016). To close off an attack vector, cyber security professionals should exhibit some awareness about how an attacker is likely to execute. Through the analysis of current and known attack approaches as well as the state of the networks, attack predictors provide relevant information to cyber security professionals (Polatidis et al., 2017; Lyons, 2014; Kott, 2014; Ramaki & Atani, 2016).