How humans take decisions: financial markets as a case of study

Abstract

Financial markets are a clear example of a choice-dealing situation in which people have to make decisions facing uncertainty and risk. This study aims to learn how people make decisions, in order to predict them, by analysing the data collected in a social experiment and using tools of information theory and statistics. It focuses on how people's own experience ifluence their next actions and what strategies are developed finding that both the market and the previous results in uence decisions with a mutual information value of 0:045 0:010 bits and 0:050 0:010 bits respectively and that these two stochastic processes add non-redundant information to each other. Besides, in the experiment, people's memory holds for only one round. Finally, the 'toy' model tested gives a 55:200 0:016% success ratio for the market's ifluence and a 53:496 0:016% for the results influence analysis.