期刊名称：Interdisciplinary Description of Complex Systems - scientific journal
印刷版ISSN：1334-4676
出版年度：2019
卷号：17
期号：4
页码：707-715
DOI：10.7906/indecs.17.4.2
语种：English
出版社：Croatian Interdisciplinary Society Provider Homepage: http://indecs.eu
摘要：Timely economic statistics is crucial for effective decision making. However, most of them are released with a lag. Thus, "nowcasting" has become widely popular in economics, and web search volume histories are already used to make predictions in various fields including IT, communications, medicine, health, business and economics. This article seeks to explore the potential of incorporating internet search data, in particular Google Trends data, in autoregressive models used to predict the volume of Bitcoin trading. Toda and Yamamoto procedure was applied in order to examine causality between Google search data and Bitcoin trading volume on the online marketplace LocalBitcoins, for the area of the Republic of Croatia. The results showed that internet search data can be useful for forecasting Bitcoin trading volume, since Google searches for the term “bitcoin” Granger causes Bitcoin trading volume in the online marketplace LocalBitcoins.
