CrowdIQ: A new opinion aggregation model

Abstract

In this study, we investigate the problem of aggregating crowd opinions for decision making. The Wisdom of Crowds (WoC) theory explains how crowd opinions should be aggregated in order to improve the performance of decision making. Crowd independence and a weighting mechanism are two important factors to crowd wisdom. However, most existing crowd opinion aggregation methods fail to build a differential weighting mechanism for identifying the expertise of individuals and appropriately accounting for crowd dependence when aggregating their judgments. We propose a new crowd opinion aggregation model, namely CrowdIQ, that has a differential weighting mechanism and accounts for individual dependence. We empirically evaluate CrowdIQ in comparison to four baseline methods using real data collected from StockTwits. The results show that, CrowdIQ significantly outperforms all baseline methods in terms of both a quadratic prediction scoring measure and simulated investment returns.