Comparing XGBoost and StockRank with Ranked Equity Factors

Summary

This note describes two stock selection approaches, XGBoost and StockRank, using the same five inputs: ranked dividend yield, market capitalization, closing price, a 20-period price ratio, and average turnover. It also gives example hyperparameter combinations for tuning each model, varying the number of leaves and trees. The author reports that StockRank performed better than XGBoost with these shared factors.

The comparison is presented as a brief implementation note, with links to the underlying strategies, rather than a full research report. It does not provide performance figures, the test period, validation design, transaction costs, or risk measures, so the reported ranking cannot establish that StockRank will generalize or outperform in other settings. The factor definitions and parameter examples may help readers reproduce the comparison, but further detail is needed to evaluate the models fairly.

Key ideas

• The two stock selection approaches use the same five ranked or transformed equity factors.
• The inputs include dividend yield, market capitalization, closing price, a 20-period price ratio, and average turnover.
• The note lists several leaves and trees settings for tuning each model.
• The author reports better results for StockRank in this comparison, without giving metrics or test details.

Tags

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