Skip to content

Alpha and research documentation

Language: English · 简体中文

This section documents reusable alpha-research interfaces and model-related methods. Project-specific hypotheses, proprietary features, and promotion decisions belong in the owning research project.

Start here

Topic Documentation
Project scope and installation Repository README
Model selection English canonical · Chinese companion
Model landscape English canonical · Chinese companion
Factor catalog English canonical · Chinese companion
Factor risk model English canonical · Chinese companion
Signal drift English canonical · Chinese companion
Factor-expression DSL English canonical · Chinese companion
Overfitting controls English canonical · Chinese companion
Research protocols English canonical · Chinese companion
Feature research English canonical · Chinese companion
Matched model and risk tools English canonical · Chinese reference
StyleReplica English canonical · Chinese companion
Signal artifact contract English canonical · Chinese companion
Research output contract English canonical · Chinese companion
AFML methods English canonical · Chinese reference
Research backends English canonical · Chinese reference
Research template English canonical · Chinese reference
Namespace migration English canonical · Chinese reference
Testing English canonical · Chinese reference
Fundamental-state forecasting English canonical · Chinese companion
Formation-date style-factor cross-sections English canonical · Chinese companion
Contextual research features English canonical · Chinese companion
Minute-factor boundaries English canonical · Chinese reference

Ownership boundary

This repository documents reusable mechanisms such as feature-evidence interfaces, model evaluation methods, and signal-artifact contracts. General portfolio backtesting, trading costs, capacity analysis, orchestration, CLI behavior, configuration composition, run directories, and target-file export are maintained here; the Python namespaces remain portfolio_backtester and strategy_pipeline.

Keep strategy hypotheses, proprietary features, model-specific portfolio rules, and promotion decisions in the owning research project. When migrating a document, update the old page with a redirect or status note so readers do not encounter multiple active copies.