Do Transformers Actually Work on Financial Time Series?
Benchmarking attention against LSTM and classical ARIMA for return forecasting. The results may surprise you.
Notes on algorithmic strategy research, backtesting discipline, data quality and quantitative thinking.
Benchmarking attention against LSTM and classical ARIMA for return forecasting. The results may surprise you.
Return-, volatility- and microstructure-based features — and how to build them without leakage.