Estimating the fractal dimension of time series of currence values using area division method

In the paper the authors propose an alternative way of assessing fractional dimension of time series. This fractional dimension, called fractal dimension, determines how the time series fills its space. It is used, e.g., to characterise series of stock exchange data, for the sake of fray degree. This gives information on how often the series is changing direction and shows whether it is a situation of return to average or a situation of trend support. The results, obtained by means of division method, are easy to interpret and link to graphic presentation of series. This is very important for practical use of this method. The authors estimated the fractal dimension for chosen time series of currency values with length of 1000 and 100 data. The method presented turned out to be very useful for both long and short series. This is another advantage of the area division method, apart from the simplicity of estimation and lack of difficulties in interpretation of results.