FinRL 强化学习交易
FinRL 强化学习交易：Use ensemble deep reinforcement learning (A2C, DDPG, PPO, TD3, SAC) to execute automated multi-market stock tr。
晶体简介
Constraints
Evidence Quality
Medium confidence — review before critical use
60 条不可违反的约束
WHENWhen implementing data download for market data acquisition
ACTIONNormalize column names to FinRL standard: date/timestamp, tic, open, high, low, close, volume
CONSEQUENCEDownstream stages (feature engineering, environment simulation) expect specific column names and will fail or produce incorrect results if columns are named differently across data sources
WHENWhen implementing data download for US equity markets
ACTIONHandle timezone as America/New_York (NYSE timezone) for each timestamp operations
CONSEQUENCETimestamp misalignment with NYSE trading hours causes incorrect alignment of price data with trading sessions, corrupting backtest results
WHENWhen downloading 1-minute interval data via yfinance API
ACTIONDownload data day-by-day in a loop due to yfinance 7-day calendar limit for intraday data
CONSEQUENCERequesting more than 7 calendar days of 1-minute data returns incomplete data or empty DataFrames, causing gaps in the training dataset
常见问题
讨论 (0)
暂无讨论，成为第一个发言的人吧！
更新历史
v0.1.0: 首次发布到 Doramagic.ai。基于 zvtvz/zvt 的自动化 batch-v1 元数据 + 自动生成 FAQ。