Distance-Weighted KNN for Multi-Indicator Trading Signals

Summary

This strategy uses a K-nearest-neighbors model to turn seven technical features into a directional prediction: price momentum, RSI, volume ratio, volatility, trend strength, MACD, and Bollinger Band position. The features are Z-score standardized so their different scales do not dominate Euclidean distance calculations. It selects nearby historical samples and weights their labels inversely by distance, then compares the resulting prediction with thresholds to open long or short positions.

The document describes a sliding historical window, probabilistic output, and example risk controls of a 2% stop loss and 4% take profit. It suggests the approach may suit trending markets with moderate volatility and sufficient liquidity, while identifying computation cost, sensitivity to K, and anomalous data as concerns. It provides no performance results. The published configuration uses hourly ETH-USDT futures data over a stated period, but the source excerpt has inconsistent parameter constraints and does not establish that the proposed market conditions or risk settings produce reliable results.

Key ideas

• The model represents each market observation with seven technical features.
• Z-score standardization is intended to make feature distances comparable.
• Closer historical samples receive greater influence through inverse-distance weighting.
• Prediction thresholds determine long and short entries, with stop and take-profit orders managing exits.
• The document notes computation, overfitting, and data-quality risks but supplies no measured performance evidence.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.