Machine-learning indicators: build a useful shortlist

A shortlist is more useful when each candidate has a distinct role. Start with the question you want to investigate, then select a few scripts whose stated methods fit it.

Examples from the recovered catalog

• Lorentzian Classification: investigate a classification-based approach.
• Adaptive SuperTrend: investigate adapting a familiar trend tool.
• Machine Learning RSI by BackQuant: compare with your existing RSI workflow.
• Machine Learning Support and Resistance: investigate how the author constructs levels.

These links identify research candidates, not independently verified recommendations. Check each current publication for its method, inputs and limitations. Names and recovered descriptions alone do not establish what a script can do.

Reduce the list

Choose one candidate and one simple baseline. Keep a record of settings, signal availability and results on an untouched period. If you cannot explain the difference, postpone adding more candidates.

Learn what AI labels mean before comparing them.