Higher values result in fewer, but higher-conviction signals.
LBank API Keys (Read-Only)
For future portfolio tracking features, you can enter your read-only LBank API keys.
Composite Sentiment Index
Signal Generation Methodology
Our signals are generated by a sophisticated, two-stage heterogeneous ensemble AI model designed to identify high-probability trading opportunities in real-time.
Stage 1: Feature Extraction
The model first processes raw market data using parallel deep learning architectures to capture different types of patterns:
Composite Sentiment Index: Aggregated scores from news, social media, and market fear/greed indices to gauge market psychology.
Volatility & Volume Analysis: Measures like Average True Range (ATR) to understand market regime and conviction.
This complete feature set is fed into a final Gradient Boosting (XGBoost) model, which acts as the "decider." It synthesizes all available evidence to produce a highly robust predictive score, which is then translated into the BUY, SELL, or HOLD signals you see in the app.
This multi-layered approach ensures our signals are not reliant on any single indicator but are the result of a comprehensive, data-driven consensus, providing a significant edge in the volatile crypto markets.