Model Architecture

At its core, the algorithm uses Convolutional Neural Networks (CNNs)—a deep learning architecture originally designed for image recognition.

IAESIR adapts this to financial data by:

• Treating price charts as visual inputs

• Recognizing patterns like head-and-shoulders, flags, double tops, and triangles

• Analyzing multi-dimensional market data:

  • Technical indicators (RSI, MACD, Bollinger Bands)

  • Order book structure (depth, liquidity walls, cluster behavior)

  • Macro signals (CPI, interest rates, economic momentum)

  • Tick-by-tick execution flow from CEXs

The system processes over 5,000 live inputs per asset, constantly scoring them to detect high-confidence trade setups.