Machine Learning: Support and Resistance

The Machine Learning: Support and Resistance indicator utilizes a K-Nearest Neighbors (KNN) inspired model to project dynamic support and resistance levels based on predictive price movement rather than static historical data.

Usage

The Usage section focuses on interpreting the dynamic zones and lines generated by the predictive model. Unlike traditional pivots, these levels adjust based on projected volatility and directional probability.

• Support and Resistance Zones: The indicator plots four primary levels. The "Strong" levels (outer lines) represent extreme projections, while the standard levels (inner lines) represent high-probability turning points. Users can identify potential reversals when price tests these zones.
• VWMA and Average Crosses: The blue line (VWMA) and purple line (Average) indicate momentum. When the VWMA is above the Average, the market is in a bullish phase; when below, it is bearish.
• Value Assessment:
  • Overvalued: Price is above the VWMA.
  • Undervalued: Price is below the VWMA.
  • Fair Value: Price is oscillating between the VWMA and the Average line.
• Long Term Memory (LTM): This level (maroon line) acts as a high-timeframe anchor for support and resistance. It updates infrequently (based on the Machine Learning Length), providing a macro perspective on price stability.

Details

The script operates by calculating a "Potential Movement" for the current bar using a source (defaulting to close), RSI, and EMA. This projection is derived from the difference between the source and its moving average over a specified lookback period.

The direction of the next bar is predicted using a multi-factor logic:

• Relationship between previous source and previous moving average.
• RSI slope and overbought/oversold crossovers (30/70 levels).
• Current price distance from the average compared to historical average distances.

The KNN component functions by populating an array with the last $X$ projections. To find specific levels, the script sorts this data:

• Strong Support/Resistance: Uses the lowest/highest values (Length of 2) from the sorted projection array.
• Standard Support/Resistance: Uses a larger subset of the sorted data (10% of the ML length) to create more stable zones.

Settings

Main Settings

• Source: The price data used as the basis for all internal calculations.
• Machine Learning Length: Determines the amount of projection data stored in the array for KNN calculations.
• Smoothing Length: The period used for calculating the RSI, EMA, and VWMA components.
• VWMA ML Projection Length: Specifies how far into the machine learning data the script should average to derive the VWMA plot.
• Long Term Memory: Controls the visualization of macro data. "Average" displays diagonal transitions, while "Hard Line" maintains horizontal levels until the next update cycle.
• Use Average Last Distance In Potential Movement: When enabled, incorporates the most recent price-to-MA distance into the projection, often increasing displacement in volatile markets.
• Show Current Projection: Toggles the visibility of the raw predictive line used to populate the ML model.

FAQ

How do I interpret the colored zones? The green zones represent predictive support areas where the model expects buying pressure, while red zones represent predictive resistance.

Why do the lines move before price reaches them? Because the script uses a predictive projection model rather than historical pivots, it adjusts support and resistance levels based on shifts in momentum and RSI before the price confirms the level.

How can I access this indicator? You can get access on the LuxAlgo Library for charting platforms like TradingView, MetaTrader (MT4/MT5), and NinjaTrader for free.