Learn Trading Indicators
Master the indicators that power the TrAIde AI Score
Kalman Filter: AI Signal Smoothing
The Kalman filter is an AI algorithm that learns market momentum and eliminates noise.
What is a Kalman Filter?
The Kalman filter is a mathematical algorithm (used in GPS, autopilots, and finance) that: 1. **Predicts** the next state based on current momentum 2. **Observes** the new raw measurement 3. **Corrects** its prediction if the actual data differs 4. **Updates velocity** to track acceleration/deceleration
Result: Smooth, confident signals that adapt to market conditions.
TrAIde's Kalman Configuration
**Process Noise (Q) = 0.0005**: - Extremely low = filter changes trajectory slowly - Perfect for weekly signals (we want slow, sustained trends) - Rejects daily noise automatically
**Measurement Noise (R) = 25.0**: - Very high = filter doesn't trust individual daily measurements - Requires consistent signals across days to update - Filters out spikes and false breakouts
How It Works
1. Raw Score calculated from RSI + MACD 2. Kalman filter smooths the raw score 3. Filter learns momentum direction (v = velocity) 4. Output: TrAIde Score (final, smoothed, confident)
Example: - Raw Day 1: +62 (spike from oversold + bullish MACD) - Kalman applies: +45 (filter skeptical of the spike) - Raw Day 2: +58 (sustained) - Kalman applies: +48 (gaining confidence) - Day 3-7: Consistent positive signals - Kalman Output: +62 (now confident in the trend)
Why This Eliminates False Signals
Traditional indicators give daily signals that flip-flop. The Kalman filter says "Prove it for a week, then I'll move."