A method for improving the quality of collective knowledge

Abstract

Collective knowledge is considered as a representative of a collective consisting of autonomous members. In the case if members’ knowledge states have to reflect some real world knowledge for example weather forecasts then the quality of collective knowledge is an important issue. The quality is measured by the difference between the collective knowledge and the real world knowledge. In this work, a method for improving the quality of collective knowledge is proposed by taking into account the number of members in a collective. For this aim, we experiment with different number of collective members using multi-dimensional vector structure to determine how the number of collective members influences the quality of collective knowledge. According to our experiments, collectives with more members will give better solutions than collectives with fewer members.