PACIS 2020 Proceedings


Public opinions and expectations on market dynamics reflected in Internet news have been considered as early warning signals for market monitoring. Although various attempts have achieved favorable early warning signals from Internet news, there still lack a rigorous approach to construct public opinion indices based on Internet news for macroeconomics monitoring and early warning. This paper proposed a framework to construct public opinion indices for market monitoring. A three-step keywords generation approach and a context-based sentiment scoring approach were designed to construct a dictionary. In empirical analysis, real estate data across China, first-tier cities, and Beijing were used to construct public opinion indexes. These indexes revealed around 5 months leading time to corresponding house price indicator. And the stability test proved that the proposed approach is independent from sample news website, thus our work is hopefully applicable to other market monitoring tasks and other macro-economic early warning scenarios.



When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.