Estimating strange attractor´s dimension in very noisy data, application to FOREX time series

To simply visualize and analyze the behavior of economic time series, a common tool is the Japanese candlestick chart which is widely used in modern technical analysis methods. What traders refer to as the candlestick patterns implicitly reveals the existence of some systematic dynamics behind most of economic time series, which can also be unveiled by modern nonlinear models and tests. This paper introduces an extension to Grassberger-Procaccia method, which computes the correlation dimension of Japanese candlesticks instead of correlation dimension of commonly used `close´ points of the economic time series. The results of this method applied on foreign exchange rates, satisfactorily show lower chaotic dimension of Japanese candlesticks series than regular Grassberger-Procaccia method applied merely on close points of these same candles