This article presents a new centrality measure. The presented measure was compared with state-of-the-art measures and proved to combines two features, low computational complexity and high precision of calculations. The measure has been mathematical analyzed, and then, experiments on real-life graphs were conducted. The mathematical analysis showed many essential properties of the proposed measure. Statistical analysis of the conducted experiment showed that measure has precision the same as state-of-the-art measures with much lower computational complexity.

Recently online social networks have grown enormously to become an essential marketing tool for many companies. Therefore, identifying the important or central or influential node is a crucial task. Centrality indicates the most central node or most influential node within a network. Centralization has proven to be a vital tool that helps to solve many problems. Besides the existence of numerous centralization measures, there is a lack of requirements for them. The main contribution of this article are postulates for centralization measures that solve this problem by enforcing intuition behind centralization. To prove the legitimacy of these postulates, a few existing measures that have fulfilled all postulates are shown. The proposed postulates are not only helpful during the formation of new measures but they also represent the perfect tool to compare two different measures.

The goal of this paper is to prove an essential impact of the independence aspect in collective prediction tasks. Due to the lack of research of the impact of individual crowd characteristics on collective prediction accuracy, we approached the problem holistically taking into account diversity, decentralization, independence and aggregation aspects of a collective. This new approach allows discovering the impact of independence in a crowd on its performance.

Nowadays, collective intelligence becomes more and more popular. Despite its high usability, many aspects of collective intelligence stay unexplored. Many companies have recognized the potential of collective intelligence and have begun using it. Prediction markets are the real life implementation of collective intelligence. The fact that prediction markets outperform experts makes it a great tool for predicting the future. In this paper, we try to answer important questions that have to be asked before the creation of a prediction market e. g. “What factors influence the prediction market error and how could this be minimized?”. This paper treats the problems more broadly. Therefore, the areas of collective intelligence that have a strong influence on prediction markets are also included in the problem analysis.