סמינר בניהול טכנולוגיה ומערכות מידע

הסמינר ינתן ע"י ד"ר שחר רייכמן,  

MITמאוניברסיטת  

Abstract: 

The analysis of large-scale data generated by the crowd has recently attracted extensive interest of both academics and practitioners. These data allow researchers to conduct real-time, large-scale monitoring of customer preferences and opinions; offer a myriad of opportunities for modeling and explaining economic phenomena; and facilitate accurate predictions.

This abundance of data also poses significant challenges to data collection and processing. Perhaps one of the most important challenges to emerge is how to determine, out of immense volumes of data, which data items should be selected in modeling a phenomenon of interest. In this paper, we offer a novel crowd-based concept to address this data selection problem. We label the concept “Crowd-Squared,” as it leverages crowds to amplify the predictive capacity of crowd-generated data.

To demonstrate the effectiveness of this concept, we developed an online word association game on a crowd marketplace platform to collect people’s associative thoughts about focal terms. We empirically tested our approach in three domains that have previously been used as test-beds for prediction: flu epidemics, the housing market, and unemployment. We demonstrate the effectiveness of this method by comparing its results to previous studies that used data selected by experts or by computational methods.