High resolution microprice estimates from limit orderbook data using hyperdimensional vector Tsetlin Machines

TL;DR

An error-correcting model for the microprice, a high-frequency estimator of future prices given higher order information of imbalances in the orderbook, and a computationally fast estimator using a recently proposed hyperdimensional vector Tsetlin machine framework are introduced.

Abstract

We propose an error-correcting model for the microprice, a high-frequency estimator of future prices given higher order information of imbalances in the orderbook. The model takes into account a current microprice estimate given the spread and best bid to ask imbalance, and adjusts the microprice based on recent dynamics of higher price rank imbalances. We introduce a computationally fast estimator using a recently proposed hyperdimensional vector Tsetlin machine framework and demonstrate empirically that this estimator can provide a robust estimate of future prices in the orderbook.

Figures (9)

• Figure 1: The bid and ask side of the orderbook with a buy matching a sell order as an event.
• Figure 2: Sequence of extracting the feature vector for encoding into a hyperdimensional vector for learning using the Tsetlin machine.
• Figure 3: Detailed sequence flow of extracting information from raw orderbook data to build the feature vector. Once the feature vector is computed, and the future price in number of ticks from current microprice is computed S steps in the future, the microprice correction is learned.
• Figure 4: Input into the TM is the encoded latest $N$ values of the sequence. This input is used to predict the next value governed by the $Q$-class TM
• Figure 5: Microprice adjustments for various spreads