QuantEdge Momentum ML [PRO]

The QuantEdge Momentum ML [PRO] indicator is a self-calibrating momentum oscillator driven by a k-Nearest Neighbors (k-NN) machine learning core. It learns asset-specific "RSI fingerprints" to identify historical bullish and bearish outcomes, providing a non-parametric alternative to static indicators like RSI or MACD.

Usage

The tool can be used for regime identification, mean-reversion setups, and trend-following signals. It functions as an oscillator that adapts its boundaries based on historical data rather than fixed levels.

• Regime Detection: Use the mid-line (zero) to determine the current market bias; values above the mid-line indicate a bullish regime, while values below suggest a bearish one.
• Mean Reversion: Look for the prediction line entering the dynamic Overbought (OB) or Oversold (OS) strips for potential reversal opportunities.
• Crossover Signals: Utilize the signal line crossovers (similar to MACD) to identify shifts in momentum.
• Filtered Signals: Monitor the "Signal Dots" which fire based on user-defined strictness levels, such as only showing crosses within OB/OS zones.

Details

The core engine utilizes a Dual-Horizon RSI feature vector, smoothing a fast and slow RSI via a Weighted Moving Average (WMA) to map short-term and structural momentum. The training sampler collects historical data points through three modes: MA Crossover, Periodic, or Hybrid.

The k-NN predictor identifies the closest historical matches in the 2D feature space to generate a prediction. This process includes a Bias Correction mechanism to ensure the oscillator remains centered even during strong trends. The dynamic bands (OB/OS) are adjusted by rolling standard deviation, allowing the indicator to tighten in low-volatility environments and expand during high volatility.

Settings

Machine Learning

• Neighbors (k): Sets the maximum number of historical samples used for prediction.
• Adaptive k: Scales the number of neighbors dynamically based on the current dataset size.
• Learning Mode: Determines the trigger for adding new samples to the memory (MA Crossover, Periodic, or Hybrid).
• Max Dataset Size: Caps the rolling window of historical memory to ensure performance.

Feature Engine

• Trend Length: Smoothing period for the underlying RSI and MA features.
• RSI Fast/Slow Period: The lookback windows for the two RSI dimensions.

Signal Configuration

• Prediction Style: Offers five visual rendering modes including Stratum, Neon, and Pulse.
• Filter Mode: Sets the strictness for Signal Dots (All Crosses, Zone Only, Mid Aligned, or Strict).
• Signal Period: The length of the WMA signal line.

FAQ

• How do I interpret the dynamic bands? Unlike fixed 80/20 RSI levels, these bands adapt to the asset's historical volatility. When the prediction line enters these shaded zones, it indicates a statistically significant momentum extreme relative to recent history.
• What is the best Learning Mode for low-volume assets? "Hybrid" or "Periodic" is recommended for fresh assets or those with infrequent moving average crossovers to ensure the dataset fills quickly enough for accurate predictions.
• How can I access the QuantEdge Momentum ML [PRO]? You can get access on the LuxAlgo Library for charting platforms like TradingView, MetaTrader (MT4/MT5), and NinjaTrader for free.