The aim of this paper is to show how machine learning can predict whether an individual is more powerful than others in the group. The crucial point here is to consider the structural position of the actors in the social networks in which they are embedded. The approach we have taken for constructing these intra-group networks is the aggregation of communication and support interactions. Our research is based on longitutional data from the Massively Multiplayer Online Game (MMOG) Travian that was collected over a 12-month period. The data includes 202,764 communication and 96,913 support interactions between players that we applied for the construction of interaction networks. We also had access to status information on a daily basis for 21,431 individual players who were members of 4,758 alliances. Methodically, we applied 10 established metrics from SNA-based team research in combination with the Random Forstest classification algorithm. Our results show that interaction networks are well suited to assign members into two groups of powerful (elite) and nonpowerful (non-elite) players. It turned out that the identification of non-elite members was much easier to accomplish than that of elite members. Regarding the application of multiplex networks, we could not confirm a higher explanatory power by using combined networks. In summary, we can say that the network patterns of elite members are clearly different from those of non-elite members. In this way, we were able to predict affiliation to each category with an accuracy (F1) of 0.88 for communication networks and 0.83 for support networks.

Multi-label classification is a variant of the classification problem where multiple labels may be assigned to each instance. Usually multi-label classification algorithms output a numerical score for each label, indicative of their relevance to a query instance. However, in many applications the desired output is a bipartition of the labels into relevant and irrelevant w.r.t the query instance. Bipartitions can be obtained from scores using various thresholding strategies, such as PCut strategy which selects relevant instances per label, and RCut strategy which selects relevant labels per instance. However, we suggest that a combination of both strategies would provide better classification performance. In this paper, we propose a fuzzy-based approach to combine PCut and RCut strategies, by converting the crisp relevance into fuzzy one, merging them linearly, and defuzzifying again. Our experiments shows that our hybrid approach indeed outperforms both strategies.

April 22, 2021, marked the 51st anniversary of Earth Day. With the growing imperativeness of environmental protection and sustainability, we want to study people’s collective attention and conversations on this themed day. What are the top-of-mind discourses and central topics about the earth? How do people feel about them, hopeful or pessimistic? How do they change over time, especially after the COVID-19 pandemic? To answer these, we extracted and quantified top frequent features, co-occurring hashtags, sentiment words, and latent sub-topics from about 300K tweets posted on the Earth Day of 2009, 2013, 2017, and 2021. The results demonstrated the longitudinal dynamics of people’s rhetoric and focus regarding protecting the earth – from resources conservation to climate changes, as well as the plummeted optimism toward environmental topics after the pandemic. The findings of our paper can help decision-makers to better assess the “voices of the people” and inform evidence-based decision-making.

Virtual teams are becoming increasingly important. Since they are digital in nature, their “trace data” enable a broad set of new research opportunities. Online Games are especially useful for studying social behavior patterns of collaborative teams. In our study we used longitudinal data from the Massively Multiplayer Online Game (MMOG) Travian collected over a 12-month period that included 4,753 teams with 18,056 individuals and their communication networks. For predicting team performance, we selected 13 SNA-based attributes frequently used in team and leadership research. Using machine learning algorithms, the added explanatory power derived from the patterns of the communication networks enabled us to achieve an adjusted R2 = 0.67 in the best fitting performance prediction model and a prediction accuracy of up to 95.3% in the classification of top performing teams.

The history of intellectuals consists of a complex web of influences and interconnections of philosophers, scientists, writers, their work, and ideas. How did these influences evolve over time? Who were the most influential scholars in a period? To answer these questions, we mined a network of influence of over 12,500 intellectuals, extracted from the Linked Open Data provider YAGO. We enriched this network with a longitudinal perspective and analyzed time-sliced projections of the complete network differentiating between within-era, inter-era, and accumulated-era networks. We thus identified various patterns of intellectuals and eras and studied their development in time. We show which scholars were most influential in different eras, and who took prominent knowledge broker roles. One essential finding is that the highest impact of an era's scholar was on their contemporaries, and that the inter-era influence of each period was strongest on the consecutive era. Furthermore, we see quantitative evidence that there was no rediscovery of Antiquity during the Renaissance; rather, there has been a continuous reception of it since the Middle Ages.

Online firestorms on Twitter are seemingly arbitrarily occurring outrages towards people, companies, media campaigns and politicians. Moral outrage can create an excessive collective aggressiveness against one single argument, one single word, or one action of a person resulting in hateful speech. With a collective “against the others” the negative dynamics often start. Using data from Twitter, we explored the starting points of several firestorm outbreaks. As a social media platform with hundreds of millions of users interacting in real-time on topics and events all over the world, Twitter serves as a social sensor for online discussions and is known for quick and often emotional disputes. The main question we pose in this article is whether we can detect the outbreak of a firestorm. Given 21 online firestorms on Twitter, the key questions regarding the anomaly detection are: 1) How can we detect changing points? 2) How can we distinguish the features that indicate a moral outrage? In this paper we examine these challenges developing a method to detect the point of change systematically spotting on linguistic cues of tweets. We are able to detect outbreaks of firestorms early and precisely only by applying linguistic cues. The results of our work can help detect negative dynamics and may have the potential for individuals, companies, and governments to mitigate hate in social media networks.

In this paper, we present an approach for classifying movie genres based on user-ratings. Our approach is based on collaborative filtering (CF), a common technique used in recommendation systems, where the similarity between movies based on user-ratings, is used to predict the genres of movies. The results of conducted experiments show that our genres classification approach outperforms many existing approaches, by achieving an F1-score of 0.70, and a hit-rate of 94%.

Linked Open Data (LOD) refers to freely available data on the WWW that are typically represented using Resource Description Framework (RDF). LOD is an invaluable source of rich and structured information, and enables a wide range of new applications, such as Social Network Analysis (SNA). In this paper, we address the extraction of social networks from LOD using SPARQL language, and we present various patterns to extract ego-centric networks. We also present two case studies: i) influence networks of intellectuals, and ii) co-acting networks, to demonstrate the applicability and usefulness of the approach.