MarketTrader AI
- Experts
- Bohdan Suvorov
- Version: 1.3
- Activations: 15
An automated trading advisor that uses artificial neural networks to predict AUDCAD price movements on the H1 timeframe. The system analyzes market data and opens positions based on bar opening prices.
Key Features:
- Currency pair: AUDCAD
- Timeframe: H1 (1 hour)
- Entry type: Bar opening price
- Position sizing: Fixed percentage of account balance
- Minimum recommended deposit: $1000
- Recommended leverage: 1:500
- Trade direction: Both long and short positions
- Optimization period: 2022-2024
Neural Network Algorithm:
The advisor implements a deep neural network with the following architecture and workflow:
1. Input Layer:
- Previous 24 hours of price data
- Opening prices of the last 24 H1 bars
- Price momentum indicators
- Volume indicators
- Market volatility metrics
2. Processing:
- Three hidden layers using ReLU activation functions
- Dropout layers to prevent overfitting
- Batch normalization for stable training
- Time-series specific attention mechanism
3. Output Layer:
- Probability of price movement direction for the next bar
- Estimated price movement magnitude
Trading Logic:
1. At the opening of each new H1 bar, the system:
- Collects and normalizes input data
- Processes it through the neural network
- Generates prediction for the next bar's direction
2. Position Entry:
- Opens long position if upward movement probability exceeds 70%
- Opens short position if downward movement probability exceeds 70%
- Position size is calculated as fixed percentage of current balance
3. Risk Management:
- Stop-loss: Dynamic, based on recent volatility
- Take-profit: Set according to predicted movement magnitude
- Maximum position size: 5% of account balance
4. Model Updates:
- Neural network retrains weekly on the latest market data
- Adaptive learning rate based on prediction accuracy
- Regular performance evaluation and parameter optimization
