NUMERICAL MARKET DATA RULES
Numerical CSV/XLSX data should be treated as structured time-series data.
Recommended metadata: symbol, timeframe, timestamp, source_file, row_id, dataset_type, feature_columns, label_columns, synthetic_timestamp.
Recognize common columns case-insensitively: timestamp/date/datetime/time, symbol/ticker, open, high, low, close, volume, vwap, atr, rsi, macd, ema, sma, adx, bbands, and other indicators.
Do not round values unless the user explicitly requests rounding. Do not convert numeric data into prose as the only representation. For numerical fine-tuning, retain structured fields and optionally add a natural-language market_state field.
Rows must be sorted by timestamp. Duplicate timestamps should be preserved only when the dataset has a valid reason, such as multiple symbols or levels.
A generated timestamp must be labeled synthetic_timestamp=true. It is a sequence index, not evidence of a real historical date.
The dataset creator should support labels such as: setup_present, setup_type, market_regime, direction, entry, stop_loss, take_profit, risk_reward, outcome, approval, denial_reason, and confidence. Labels must come from the source data or user-provided annotations.
Xet Storage Details
- Size:
- 1.23 kB
- Xet hash:
- 821dc7d434d028bf33be15fec75d5382bb909d16370a24e2390f12eea11d580e
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.