SARCP - Exploiting Cyber-Attack Prediction Through Socially-Aware Recommendation

In the domain of cyber security, the defence mechanisms of networks has traditionally been placed in a reactionary role. Cyber security professionals are therefore disadvantaged in a cyber-attack situation due to the fact that it is vital that they maneuver such attacks before the network is totally compromised. In this paper, we utilize the Betweenness Centrality network measure (social property) to discover possible cyber-attack paths and then employ computation of similar personality of nodes/users to generate predictions about possible attacks within the network. Our method proposes a social recommender algorithm called socially-aware recommendation of cyber-attack paths (SARCP), as an attack predictor in the cyber security defence domain. In a social network, SARCP exploits and delivers all possible paths which can result in cyber-attacks. Using a real-world dataset and relevant evaluation metrics, experimental results in the paper show that our proposed method is favorable and effective.

Genre Familiarity Correlation-Based Recommender Algorithm for New User Cold Start Problem

The advancement of web services paved the way to the accumulation of a tremendous amount of information into the world wide web. The huge pile of information makes it hard for the user to get the required information at the right time. Therefore, to get the right item, recommender systems are emphasized. Recommender algorithms generally act on the user information to render recommendations. In this scenario, when a new user enters the system, it fails in rendering recommendation due to unavailability of user information, resulting in a new user problem. So, in this paper, a movie recommender algorithm is constructed to address the prevailing new user cold start problem by utilizing only movie genres. Unlike other techniques, in the proposed work, familiarity of each movie genre is considered to compute the genre significance value. Based on genre significance value, genre similarity is correlated to render recommendations to a new user. The evaluation of the proposed recommender algorithm on real-world datasets shows that the algorithm performs better than the other similar approaches.

LDA-LFM

Most of the existing recommender systems are based only on the rating data, and they ignore other sources of information that might increase the quality of recommendations, such as textual reviews, or user and item characteristics. Moreover, the majority of those systems are applicable only on small datasets (with thousands of observations) and are unable to handle large datasets (with millions of observations). We propose a recommender algorithm that combines a rating modeling technique (i.e., Latent Factor Model) with a topic modeling method based on textual reviews (i.e., Latent Dirichlet Allocation), and we extend the algorithm such that it allows adding extra user- and item-specific information to the system. We evaluate the performance of the algorithm using Amazon.com datasets with different sizes, corresponding to 23 product categories. After comparing the built model to four other models, we found that combining textual reviews with ratings leads to better recommendations. Moreover, we found that adding extra user and item features to the model increases its prediction accuracy, which is especially true for medium and large datasets.

Recommender system on social networking site with domain specific and sparse data

In many recent domain-specific social networking sites, posts are organized in chronological order, where later posts are shown first at the top, even though they might not be of everyone's interest. As a result, if users want to read posts that interest them, they will have to scroll down and sift through all the posts. To overcome this information overload problem and relieve users' burden, a recommender system is needed in social networking sites. In this thesis we propose a hybrid approach of Recommender System (RS) that combines both Collaborative Filtering and Content-based approach. Although each approach has their own weaknesses independently, by joining them together we can improve the accuracy of our recommendations. From our experiments, we noticed that using learning to rank algorithms in combining each recommender algorithm greatly enhances the system's performance.

Recommender system on social networking site with domain specific and sparse data

In many recent domain-specific social networking sites, posts are organized in chronological order, where later posts are shown first at the top, even though they might not be of everyone's interest. As a result, if users want to read posts that interest them, they will have to scroll down and sift through all the posts. To overcome this information overload problem and relieve users' burden, a recommender system is needed in social networking sites. In this thesis we propose a hybrid approach of Recommender System (RS) that combines both Collaborative Filtering and Content-based approach. Although each approach has their own weaknesses independently, by joining them together we can improve the accuracy of our recommendations. From our experiments, we noticed that using learning to rank algorithms in combining each recommender algorithm greatly enhances the system's performance.

Location-based Orientation Context Dependent Recommender System for Users

As the technology revolving around IoT sensors develops in a rapid manner, the subsequent social networks that are essential for the growth of the system will be utilized as a means to filter the objects that are preferred by the consumers. The ultimate purpose of the system is to give the customers personalized recommendations based on their preference. Similarly, the location and orientation will also play a crucial role in identifying the preference of the customer is a more efficient manner. Almost all social networks make use of location information to provide better services to the users based on the research performed. Hence there is a need for developing a recommender system that is dependent on location. In this paper, we have incorporated a recommender system that makes use of recommender algorithm that is personalized to take into consideration the context of the user. The preference of the user is analysed with the help of IoT smart devices like the smart watches, Google home, smart phones, ipads etc. The user preferences are obtained from these devices and will enable the recommender system to gauge the best resources. The results based on evaluation are compared with that of the content-based recommender algorithm and collaborative filtering to enable the recommendation engine’s power.

Evaluating Platform Accountability: Terrorist Content on YouTube

YouTube has traditionally been singled out as particularly influential in the spreading of ISIS content. However, the platform along with Facebook, Twitter, and Microsoft jointly created the Global Internet Forum to Counter Terrorism in 2017 as one mode to be more accountable and take measures toward combating extremist content online. Though extreme content on YouTube has been found to have decreased substantially due to this and other efforts (human and machine-based), it is valuable to historically review what role YouTube previously had in order to better understand the evolution of contemporary moves toward platform accountability in terms of extremist video content sharing. Therefore, this study explores what role YouTube’s recommender algorithm had in directing users to ISIS-related content prior to large-scale pressure by citizens and governments to more aggressively moderate extremist content. To investigate this, a YouTube video network from 2016 consisting of 15,021 videos (nodes) and 190,087 recommendations between them (edges) was studied. Using Qualitative Comparative Analysis, this study evaluates 11 video attributes (such as genre, language, and radical keywords) and identifies sets of attributes that were found to potentially be involved in the outcomes of YouTube recommending extreme content. This historical review of YouTube at a unique point in platform accountability ultimately raises questions of how platforms might be able to be more proactive rather than reactive regarding filtering and moderating extremist content.

Genetic Algorithm Influenced Top-N Recommender System to Alleviate New User Cold Start Problem

Most recommender systems are based on the familiar collaborative filtering algorithm to suggest items. Quite often, collaborative filtering algorithm fails in generating recommendations due to the lack of adequate user information resulting in new user cold start problem. The cold start problem is one among the prevailing issue in recommendation system where the system fails to render recommendations. To overcome the new user cold start issue, demographical information of the user is utilised as the user information source. Among the demographical information, the impact of the user gender is less explored when compared with other information like age, profession, region, etc. In this work, a genetic algorithm-influenced gender-based top-n recommender algorithm is proposed to address the new user cold start problem. The algorithm utilises the evolution concepts of the genetic algorithm to render top-n recommendations to a new user. The evaluation of the proposed algorithm using real world datasets proved that the algorithm has a better efficiency than the state of art approaches.

Genetic Algorithm-Influenced Top-N Recommender System to Alleviate the New User Cold Start Problem

Most recommender systems are based on the familiar collaborative filtering algorithm to suggest items. Quite often, collaborative filtering algorithm fails in generating recommendations due to the lack of adequate user information resulting in new user cold start problem. Cold start problem is one of the prevailing issues in recommendation system where the system fails to render recommendation. To overcome the new user cold start issue, demographical information of the user is utilised as the user information source. Among the demographical information, the impact of user gender is less explored when compared with other information like age, profession, region, etc. In this chapter, genetic algorithm influenced gender-based top-n recommender algorithm is proposed to address the new user cold start problem. The algorithm utilises the evolution concepts of genetic algorithm to render top-n recommendations to a new user. The evaluation of the proposed algorithm using real world datasets proved that the algorithm has a better efficiency than the state-of-art approaches.