Meronexa processes order book depth, volume profiles and volatility clusters in parallel and delivers quantified signals with understandable justification - without black box logic.
Meronexa's infrastructure is designed for parallel processing, not sequential queries. This significantly reduces the time between market entry and signal output compared to classic polling approaches.
Example display of the live feed in the terminal, no real-time data on this page.
Each data point goes through a normalization layer before being included in pattern recognition. This ensures that comparisons between crypto, foreign exchange and index markets remain consistent, even if the output formats of each data provider are structured differently.
Meronexa was developed for traders who want to make decisions based on understandable data instead of subjective assessment. The system combines statistical models with rule-based validation so that every signal remains traceable to a specific trigger.
The development focuses on three points: low latency, comprehensible model logic and a database that is large enough to distinguish statistically reliable patterns from short-term market noise.
The model combines time series analysis with a rule-based validation layer. Statistical abnormalities – such as unusual order book asymmetries or volatility clusters – are first identified and then checked against historical comparative patterns. Only matches above a defined confidence level generate a signal.
This two-stage test reduces false signals, which are often caused by short-term noise in pure pattern recognition models.
Workflow
Each signal can be traced back to the original data point. This distinguishes the approach from models that make decisions exclusively via non-transparent neural weights.
Real-time data streams from order books, trading volumes and reference prices are continuously recorded and checked for completeness. Gaps or delayed data points are marked, not artificially interpolated.
Statistical models identify deviations from historical distributions, such as volume peaks or volatility clusters. The weighting of the individual factors is documented and visible.
Each recognized pattern is checked against fixed rules, such as minimum liquidity or maximum slippage tolerance. Only after this check does a signal reach the terminal.
The weighting of the model parameters differs depending on the time horizon. The following overview shows which module is relevant for which approach.
| Trading style | Time horizon | Data focus | Relevant module |
|---|---|---|---|
| Scalping | Seconds to minutes | Order book depth, tick volume | Latency optimized signal pipeline |
| Swing trading | Hours to days | Volatility clusters, trend strength | Medium-term pattern recognition |
| Portfolio hedging | days to weeks | Correlation matrix, drawdown risk | Risk management parameters |
Meronexa provides a REST interface and a WebSocket feed for real-time data. Both interfaces are documented and can be connected to existing order management systems without revealing the internal infrastructure.
The statistical basic models are readjusted at regular cycles with current market data. Validation rules are adjusted independently as soon as structural market conditions, such as liquidity regimes, demonstrably change.
Connections are transmitted encrypted, access to account and configuration data is restricted based on roles. Market data and account information are processed separately and used exclusively for signal generation.
Meronexa is aimed at traders and analysts who work with structured data. Access is staggered into three levels, depending on data depth and API quota.
Access to the signal feed and basic dashboards, limited API quota.
Full signal feed, advanced risk parameters, higher API quota.
Individual integration, dedicated data quotas, extended support.
