Abstract
Due to the developments in information technology, it has become possible to utilize the Internet and websites in order to complement the incompleteness of market functionality. In this backdrop, we have studied how to design a web council system, a web-based social optimization technique. This paper points out spam-voting as issues involving the web council system. In addition, we propose a stratified estimation method as a learning method that will improve the reliability of predicting the results of voting on the basis of a straw poll. This study has verified and calculated the results of the stratified estimation method against those of simple estimation and Bayesian estimation. Community functionality provided by the Internet should be utilized to complement market functionality and build a national consensus.
