Transformer for Times Series: an Application to the S&P500

By Pierre Brugière, Gabriel Turinici

Rating

1485
Battle Count: 101

Relevance

7/10
Directly applicable to financial time series prediction, particularly for volatility forecasting

Implementation Complexity

6/10
Requires understanding of transformer architecture and adaptation for time series data

Reproducibility

3/5
Code structure and parameters are described, but full implementation details are not provided

About this paper

Methodology: Transformer Encoder. Problem types: Time Series Forecasting, Classification.

The interactive Everscope explorer (charts, battles, favorites) loads below.