Add a cross-market public-data validation target

[initial-d]

Context: the Awesome AI Trading Research full-text review included this caveat:

Real-data validation is limited to 3 years (2022-2024) of A-shares only; cross-market generalization would add confidence.

Goal: define and run one additional public-data validation target outside the current A-share focus, while preserving the same caveats around data quality, costs, and non-investment use.

Suggested scope:

• choose a small, redistributable or on-demand public-data universe, such as US ETFs/equities through yfinance;
• predeclare the universe, date range, data source, cost assumptions, and failure modes;
• run a compact validation or factor-IC workflow that fits normal contributor machines;
• publish the exact command, generated artifacts, and caveats;
• compare the result to the A-share public validation only as a scope check, not as a market ranking.

Useful entry points:

• docs/public_data_mini_reproduction.md
• docs/public_data_validation.md
• scripts/public_data_validation.py
• scripts/eval_factor_ic.py
• review comment: Paper suggestion: Machine Learning Enhanced Multi-Factor Quantitative Trading ohselab/awesome-ai-trading-research#1 (comment)

Acceptance criteria:

• The new market/universe is documented before interpreting results.
• The workflow can be reproduced without proprietary data.
• The report includes costs, turnover or factor-IC caveats, and data-source limitations.