Identification of high-frequency traders using fuzzy logic methods

Nowadays there is no single definition of high-frequency traders (HFT). In this paper HFT were classified as traders that have low value of the average of 10% of shortest position holding times, low value of the average of the ratios of end-of-day position to volume traded and high ratio of canceled orders to submitted orders. For these ratios, the fuzzy membership functions were constructed based on the analysis of the related literature. Then, for various cutoffs of membership functions, traders were classified as HFT or non-HFT. For given set of features and membership functions the share of HFT volume traded at USDRUB_TOM at MOEX FX Market (March-May 2017) was estimated between 40% and 60%.