Chinese Equity AI Stock Selection: Weekly Model Performance Review

Summary

This report compares artificial intelligence stock-selection portfolios across five Chinese equity universes and benchmark or sector-neutral setups. The models include support vector machines, random forests, naive Bayes, and logistic regression. It reports weekly, recent three-month, and one-year absolute and excess returns against the CSI 300 or CSI 500, depending on the selection universe. SVM leads several weekly and three-month comparisons, while random forests and naive Bayes lead some one-year or sector-neutral comparisons; no single model dominates every setup.

The figures are historical snapshots from a report dated May 13, 2018, and the text does not describe model features, training procedures, portfolio construction, trading costs, or statistical tests. Some return fields are missing or incomplete, which limits comparisons. The report explicitly warns that model-based selection reflects past experience and may stop working. Its results therefore document relative historical outcomes across the listed configurations, rather than establishing that any model will continue to outperform.

Key ideas

• The report compares several machine-learning stock-selection models across broad-market, sector-neutral, and index-constituent universes.
• SVM leads multiple weekly and three-month comparisons, while other models lead some one-year results.
• Performance is reported against CSI 300 or CSI 500 benchmarks, depending on the portfolio setup.
• Several reported return fields are incomplete, and the document gives little detail on model design or trading costs.
• The report warns that historical model behavior may fail to persist.

Tags

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