TabNet Stock Selection Strategy Using Price and Volume Factors

Summary

This short strategy summary describes a quantitative stock-selection approach built with TabNet, a machine-learning model. It says the approach extracts 98 price and volume factors, trains the model on data from 2010 through 2018, and applies predictions to a backtest covering 2018 through September 2021. The stated workflow is therefore to use historical market-derived factors as model inputs and evaluate selections on a later period.

The source provides no details about the target variable, factor definitions, portfolio construction, rebalance schedule, transaction costs, or benchmark. It also reports no returns or risk statistics, so the existence of a backtest cannot establish that the strategy was profitable or robust. The brief entry is useful as a high-level outline of the model and date split, but further documentation would be required to reproduce or assess the approach.

Key ideas

• The strategy uses TabNet to produce stock-selection predictions.
• It draws on 98 price and volume factors.
• The model is trained on data from 2010 through 2018 and tested on a later backtest period through September 2021.
• The source omits factor definitions, portfolio rules, costs, and performance statistics.

Tags

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