Machine Learning Price Target Prediction Signals

The Machine Learning Price Target Prediction Signals tool leverages non-parametric kernel regression to estimate potential price moves and generate automated take-profit and stop-loss levels based on historical feature correlations. By analyzing volatility, relative strength, and trend direction, the indicator provides data-driven targets to assist traders in optimizing their risk management.

Usage

The script identifies trend changes using one of three selectable methods and then projects a price target based on learned patterns. To use the tool:

• Trend Detection Selection: Choose between SuperTrend, Hull Moving Average (HMA), or EMA Crossovers to define the core signal logic.
• Analyzing Predictions: When a signal appears, the tool plots dynamic target boxes. The "Predicted Move" indicates the expected magnitude of the price change, while the "Recommended RR" adapts based on the historical performance of the model.
• Risk Management: Use the green and red visual trade levels for setting take-profit and stop-loss orders. The indicator automatically tracks these levels and colors the bars to indicate if a trade is currently "active" or has hit a target.
• Performance Tracking: An on-chart dashboard displays the real-time win rate and training data size, allowing you to gauge the reliability of the current predictions.

Details

The core of the indicator is a Kernel Regression model (using a Gaussian RBF kernel). Unlike linear models, kernel regression weights historical data points based on their similarity to the current market state.

The model evaluates eight distinct features to make a prediction:

1. Dump Intensity: Measures the speed of price drops from recent highs.
2. Pump Intensity: Measures the speed of price surges from recent lows.
3. Volatility Delta: Calculates the deviation of current volatility from its moving average.
4. Acceleration of Volatility: The rate of change in volatility movements.
5. Volume Oscillator: Measures volume momentum.
6. Choppiness Index: Determines if the market is trending or ranging.
7. Relative Strength Index (RSI): Standard momentum measurement.
8. Trend Bias: The direction of the underlying trend filter.

When a trend reversal occurs, the script looks back at the "absolute move" achieved during previous trends that shared similar feature profiles to predict the current move's potential.

Settings

Trend Indicator Settings

• Trend Indicator for Signals: Select the logic used for entry (SuperTrend, HMA, or EMA Cross).
• Hull Moving Average Length: The lookback period for HMA trend detection.
• Fast/Slow EMA Length: Periods for the EMA crossover logic.
• SuperTrend Factor/ATR Length: Parameters defining the sensitivity of the SuperTrend filter.

Kernel Regression Settings

• Kernel Regression Bandwidth: Controls the weight of distant data points. Higher values result in smoother, more generalized predictions, while lower values make the model more sensitive to recent local data.
• Hide NaN Predictions: Removes signals where internal calculations do not produce a valid numerical prediction.

Visuals

• Bullish/Bearish Color: Customizes the color scheme for labels, boxes, and bar coloring.

FAQ

How do I interpret the "Predicted Move" value? The predicted move represents the percentage change the model expects the price to undergo before a trend reversal or target hit, based on its training data.

Why are some signals hidden? If "Hide NaN Predictions" is enabled, the script will filter out signals where the kernel regression lacks sufficient similar historical data to make a confident prediction.

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.