Lorentzian Classification

Category: trend, momentum

A machine learning classifier that uses Lorentzian Distance to find historical similarities in multi-dimensional feature space for price prediction.

Formula

d(x,y) = \sum_{i=1}^{n} \ln(1 + |x_i - y_i|) \text{ followed by K-Nearest Neighbors voting}

Inputs

• Neighbors Count (default: 8)
• Max Bars Back (default: 2000)
• Feature Count (default: 5)
• Feature Selection (default: RSI, WT, CCI, ADX)

See signal primitives and every published strategy that uses Lorentzian Classification on WOBR StrategyVerse.