Impact of Discretization Units on the Modelling Efficiency of Currency Exchange Rate in Binary Representation

Exchange rate course trajectory can be visualised in a binary representation. Binarization algorithm involves transformation of a course represented by tic data into a respective binary string. Each exchange rate increase by one discretization unit is assigned a binary value 1, and a respective decrease is assigned a 0. Properties and character of the exchange rate course in binary representation, that is, for example, influence of random fluctuations (noise) or relationship between current changes and historical data, are dependent on the appointed discretization unit. In the article a statistical analysis of an exchange rate trajectory for AUD/NZD currency pair in binary representation with different discretization units is presented. The analysis was performed in order to verify possible application of binary representation to create prediction systems based on binary representation state model (BRSM). Data used in performed research consisted of tick data for AUD/NZD from a five-year time period 2012–2017.