Prediction Market System using VIPS, Gaussian Forecast Distribution and Batch-updating Central Market Maker for Demand Forecasting

抄録

This paper presents an intra-company prediction market system as a collective-knowledge-based demand forecasting tool. The system uses the variable interval prediction security (VIPS) to be traded in the market and is controlled by a newly introduced computerized central market maker (CMM). The CMM has a Gaussian forecast distribution and updates the distribution on a batch basis through an inventory-based updating logic according to the transactions of the VIPS in the market. An agent-based simulation and laboratory experiments are performed to confirm whether the proposed system functions properly as a subjective forecasting tool.