Explanation of Binarized Time Series using Genetic Learning Model of Investor Sentiment

Abstract

The aim of this paper is to reveal the relations between time scales and time series properties by concentrating on information requisite for speculators using a genetic learning model of investor sentiment. For this purpose, first the authors identified the conditions for describing investor sentiment by altering parameters of genetic algorithm. Then auto-correlations and conditional probabilities were calculated using the estimated models in the first step. The results show that both the amount and quality of information for the agents determine the time series properties. This implies that the preciseness of information which speculators permit depends on their time scales.