Skip to content
All library documents

Deep Belief Networks for Treasury Futures Portfolio Trading

Article arXiv papers · Author: Abhijit Sharang et al.
Read the original

Summary

This study describes a medium-frequency strategy for a portfolio of five-year and ten-year US Treasury note futures. It frames trading as a classification task: predict whether the portfolio will move up or down over the coming week. The prediction inputs are features learned by a deep belief network trained on technical indicators from the portfolio’s constituent futures.

The authors report that the resulting machine-learning pipeline produced profitable trades in their experiments. The available description gives no details about the sample period, test design, costs, risk-adjusted performance, or comparison with simpler models. As a result, it conveys the modeling approach but provides limited evidence for judging how robust or deployable the reported profitability is. Historical experimental success should not be read as assurance of future returns.

Key ideas

  • The strategy trades a portfolio of five-year and ten-year US Treasury note futures.
  • It predicts the portfolio’s weekly direction as a classification problem.
  • A deep belief network extracts features from technical indicators of the futures constituents.
  • The authors report profitable experimental trading results.
  • The description does not provide enough detail to assess robustness or trading costs.

Tags

FuturesFixed incomeMachine learningTechnical indicators
Full text
# Using machine learning for medium frequency derivative portfolio trading


# Using machine learning for medium frequency derivative portfolio trading









We use machine learning for designing a medium frequency trading strategy for a portfolio of 5 year and 10 year US Treasury note futures. We formulate this as a classification problem where we predict the weekly direction of movement of the portfolio using features extracted from a deep belief network trained on technical indicators of the portfolio constituents. The experimentation shows that the resulting pipeline is effective in making a profitable trade.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.