Dynamic parameters for Brazilian financial time series

Abstract

In this study we analyse a Brazilian stock index called IBOVESPA using techniques from dynamical systems theory and stochastic processes. We discuss the Lyapunov exponent, the correlation dimension, the LempelZiv complexity, the Hurst exponent and the BDS statistics. We compare this study with other time series including stock prices and deterministic systems. We conclude that the IBOVESPA is a linear stochastic process that exhibits the phenomenon of persistence, thatis, it has long term memory. The stocks are described by nonlinear stochastic processes making it impossible to be simulated with deterministic models such as the usual neural networks architectures.