Rewards Visualization System Promotes Information Provision

Abstract

This is an extension from a selected paper from JSAI2019. If users participated in a Consumer Generated Media (CGM) sites on the web do not provide sufficient information to the site, the site would be insignificant. Despite this point, many users have lower incentives to provide something than those to receive something. This is why such the information-sharing sites face the so-called free-riding issues. Therefore, the information behaviors of CGM are regarded as a kind of a public good game. In this paper, we model such behaviors to a restricted meta-reward game and analyze some effects of their information behaviors on CGM to be activated using agent-based simulations. Our simulation shows that CGM can be activated if at least half agents observe the other agents’ meta-rewards information. Moreover, our results show that agents evolutionarily refer individual meta-reward rates to keep cooperation when they can learn their reference strategies.