ARTIFICIAL INTELLIGENCE AND ECONOMIC GROWTH

Shigeyuki Hamori and Takahiro Kume
Additional contact information
Takahiro Kume: Graduate School of Economics, Kobe University, Kobe, Japan

Advances in Decision Sciences, 2018, vol. 22, issue 1, 256-278

Abstract: This paper describes the use of five machine learning methods for predicting economic growth based on a country’s attributes and presents a comparison of their prediction accuracy. The methods used are four neural network (NN) methods with different activation functions, and eXtreme Gradient Boosting (XGBoost). Their performance is compared in terms of their ability to predict the economic growth rate using three measures (prediction accuracy rate, area under the curve (AUC) value, and F-score). The results obtained can be summarized as follows: 1) XGBoost outperforms the NNs in terms of prediction accuracy and F-score for original data; 2) data standardization enhances the reliability of NNs, improving their prediction accuracy, AUC-value, and F-score; 3) XGBoost has smaller standard deviation of prediction accuracy rate than that of NNs; and 4) "Political institution", "Investment and its composition", "Colonial history", and "Trade" are important factors for cross-country economic growth.

Keywords: Economic growth; machine learning; XGBoost; neural network (search for similar items in EconPapers)
JEL-codes: C45 E10 (search for similar items in EconPapers)
Date: 2018
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://iads.site/Hamori_Kume__ADS (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:aag:wpaper:v:22:y:2018:i:1:p:256-278

Access Statistics for this article

More articles in Advances in Decision Sciences from Asia University, Taiwan Contact information at EDIRC.
Bibliographic data for series maintained by Vincent Pan ().