HomeAbstract of 4th-bmeconfForecasting of Important Economic Indicators based on Mobile Data and its application to effective trading

Forecasting of Important Economic Indicators based on Mobile Data and its application to effective trading

Proceedings of The 4th International Conference on Applied Research in Business, Management and Economics

Year: 2022

DOI:

[Fulltext]

Forecasting of Important Economic Indicators based on Mobile Data and its application to effective trading

Takuya Kaneko, Yutaro Mishima, Shinya Wada and Rui Kimura

 

ABSTRACT: 

In this paper, we propose effective forecasting method of important economic indicators by utilizing mobile data and show its application to effective trading. There are so many articles about economic indicators forecasting models and researchers tried to create them based on historical data (time series data). Their methodologies are multivariate regression or autoregression models. Our approach is completely different from them because of our utilizing mobile data and not using historical data. More specifically, we count the number of mobile phone users, who permitted us to analyse, mesh by mesh mobile spatial statistics (MSS) and utilized these statistics to estimate operating status of production area. This approach enables us to stably forecast even when its trend suddenly changed from the historical data. In this paper, we focus on Japan’s production index which is one of the most important indexes to well predict Japan’s GDP (Gross Domestic Product) and its growth rate. Production index can explain around 20% of Japan’s GDP also can explain 40% if it includes closely related industries. Many investors/economists/analysts are forecasting production index for market participants’ effective trading and investors are reallocating their trading portfolio accordingly. Production index is reported monthly by government. Preliminary report is released after one month (Ex. Production index in January is reported around the end of March) and Confirmation report is released after two months (around the end of April). Our approach enables us to forecast after a few days (Ex. Production index in January can be checked around February 3rd). In this paper, we mainly explain how to utilize MSS for forecasting production index and show numerical result (The correlation between GDP and our forecast was greater than 0.83) also introduce how to utilize them for effective trading.

keywords: Big data, GDP prediction model, Mobile Spatial Statistics, Production Index.