Exploring the Advantages of Transformers for High-Frequency Trading

This paper explores the use of deep learning Transformers architectures for high-frequency Bitcoin-USDT log-return forecasting and compares them to traditional Long Short-Term Memory (LSTM) models. The authors introduce a hybrid Transformer model, called HFformer, which incorporates a Transformer encoder, linear decoder, spiking activations, and quantile loss function without using position encoding. They also discuss possible high-frequency trading strategies for use with the HFformer model, including trade sizing, trading signal aggregation, and minimal trading threshold. The paper suggests several future lines of research that can be undertaken to improve the LOB snapshot pre-processing pipeline's noise reduction by using autoencoders with automated feature selection. Additionally, they suggest performing more extensive performance assessments of the HFformer on large forecast horizons and using altcoin trading pairs such as ETH-USDT. The authors also recommend implementing the HFformer with other types of Attention modules such as auto-correlation Attention. Moreover, they propose implementing a more realistic backtesting environment that accounts for the impact of placed orders and emulates other market participants' activity to assess the performance of the HFformer. When testing LSTM and HFformer models in log-returns forecasting from 1 to 30 ticks ahead, they achieved higher R2 scores than other deep learning architectures. Moreover, both LSTM and HFformer models achieved similar performance for classification tasks. Finally, when backtested on different trading strategies involving 1-5 trade signals during multiple signals' trades in BTC-USDT LOB data collected over two days and a month after training and validation data were found that using more than one trade signal decreases the number of trades and increases the cumulative PnL of a long-short trading strategy. The HFformer generates long and short trade signals that result in a more balanced trading strategy than LSTM when complemented with trade sizing to improve cumulative PnL. However these methods may yield different results when trading another cryptocurrency pair or financial asset as machine learning methods are data driven and sometimes less generalizable than traditional statistical methods.

Exploring Deep Learning Transformers for High-Frequency Bitcoin-USDT Log-Return Forecasting

In recent years, the cryptocurrency market has seen a surge in popularity and trading activity. As such, there is an increasing need for accurate forecasting models that can help traders make better decisions when investing in cryptocurrencies. This paper explores the use of deep learning Transformers architectures for high-frequency Bitcoin-USDT log-return forecasting and compares them to traditional Long Short-Term Memory (LSTM) models.

Introduction

The authors introduce a hybrid Transformer model, called HFformer, which incorporates a Transformer encoder, linear decoder, spiking activations, and quantile loss function without using position encoding. They also discuss possible high-frequency trading strategies for use with the HFformer model, including trade sizing, trading signal aggregation, and minimal trading threshold. The paper suggests several future lines of research that can be undertaken to improve the LOB snapshot pre-processing pipeline's noise reduction by using autoencoders with automated feature selection. Additionally, they suggest performing more extensive performance assessments of the HFformer on large forecast horizons and using altcoin trading pairs such as ETH-USDT. The authors also recommend implementing the HFformer with other types of Attention modules such as auto-correlation Attention. Moreover, they propose implementing a more realistic backtesting environment that accounts for the impact of placed orders and emulates other market participants' activity to assess the performance of the HFformer.

Log Return Forecasting Performance Assessment

When testing LSTM and HFformer models in log returns forecasting from 1 to 30 ticks ahead ,they achieved higher R2 scores than other deep learning architectures .Moreover both LSTM and HF former models achieved similar performance for classification tasks .

Backtesting Results

When backtested on different trading strategies involving 1 - 5 trade signals during multiple signals' trades in BTC - USDT LOB data collected over two days and month after training validation data were found that using more than one trade signal decreases number of trades increases cumulative PnL long short trading strategy .The HFFormer generates long short trade signals results balanced trading strategy than LSTM when complemented with trade sizing improve cumulative PnL .However these methods may yield different results when another cryptocurrency pair or financial asset as machine learning methods are data driven sometimes less generalizable than traditional statistical methods .

Conclusion

This paper explored how deep learning Transformers architectures could be used to accurately forecast high frequency Bitcoin - USDT log return values compared to traditional Long Short Term Memory (LSTM) models .The authors introduced hybrid Transfomer model called HFFormer which incorporated transformer encoder ,linear decoder ,spiking activations ,quantile loss function without position encoding discussed possible high frequency strategies use HFFormer model including trade sizing ,trading signal aggregation minimal thresholds .When tested against different backtesting scenarios showed improved accuracy over existing deep learning architectures while maintaining similar classification task performances .Finally proposed implementation HFFormer other attention modules auto correlation attention more realistic backtesting environment account impact placed orders emulate market participants activity assess performance HFFormer further research needed explore noise reduction autoencoders automated feature selection larger forecast horizons altcoin pairs ETH - USDT order understand true potential this architecture applied cryptocurrency markets