Skip to content

features.md

Latest commit

 

History

History
19 lines (12 loc) · 1.15 KB

features.md

File metadata and controls

19 lines (12 loc) · 1.15 KB

TradePulse Features

Real-Time Feature Store

The feature store holds the current feature vector per ticker: VWAP (5 min), volume z-score, price momentum (1 min), trade frequency, and optional bid-ask spread. Partition key is ticker#YYYY-MM-DD-HH (hour bucket) for write distribution. TTL 7 days. Used by the API (GET /features/{ticker}) and by anomaly detection as inputs. Features are updated on every validated trade in the Faust pipeline.

Aggregations

  • VWAP (1 min and 5 min): Volume-weighted average price; tumbling windows.
  • Rolling average price: 5-minute hopping window (hop 1 min).
  • Volume z-score: (volume - mean_20) / std_20 over last 20 periods.
  • Price momentum: (current_price - price_60s_ago) / price_60s_ago * 100.
  • Trade frequency: Trades per minute in current window.

Stored in DynamoDB market_aggregations (pk = ticker, sort_key = window_start).

Anomaly Detection

Per-ticker Isolation Forest with rolling training window. Contamination 0.01. Features: price, volume, volume_zscore, price_momentum, vwap_deviation, trade_frequency. Retrain every N events. Anomalies written to DynamoDB and S3.