An Intelligent Agent to Classify Countries Based on Financial Indices

Traditionally, the countries classification is performed based on several features, that are related to economic and social factors. However, the classification process is costly due to the difficulty of obtaining those features and the need for intervention of human expertise. In this paper, we propose an intelligent agent that classifies countries based on financial indices. The artificial agent calculates the probability density function (pdf) of the return series of financial indices. This pdf characterizes the fluctuation of a market. Based on the return series and pdf, the volatility and the B coefficient of the exponential function, that describe the behavior of world markets, are estimated. Then, the intelligent agent classifies the indices from developed and developing countries using a Self-Organizing Map (SOM) neural network. The results show that the proposed intelligent agent is an accurate, fast and cheap alternative to the classifications provided by traditional organizations.