An Independence Measure for Expert Collections Based on Social Media Profiles

Abstract

According to current research, a crowd can outperform experts. Surowiecki in his work, has distinguished decentralization, independence, and diversity as key factors of good crowd performance. Due to lack of mathematical models for modelling these aspects, it is still impossible to prove that they have a big impact on the crowd performance. For solving this problem one of the very important crowd metrics called independence measure should be defined and this is the objective of our paper. Proposed measure allows calculating independence values based on data from social media profiles. The biggest advantage of the measure is the possibility of calculating an independence value for a group of people before it could become a collective for realizing concrete objectives. Currently, known solutions largely simplify the problem by describing independence with a single value. The solution presented in the article assumes that the value of the independence is calculated for a specific topic (the calculated value is part of a vector describing the independence between two experts).