D'où proviennent les données utilisées par les modèles ?

Les modèles s'appuient sur des flux de prix, de volumes et d'indicateurs de liquidité provenant de plusieurs plateformes d'échange. Ces flux sont agrégés et vérifiés avant intégration dans les modèles prédictifs.

Comment la sécurité des actifs est-elle assurée ?

Virelo Xanum produit des recommandations d'allocation et d'ajustement de risque, mais ne détient pas directement les actifs des utilisateurs. La conservation reste sous la responsabilité de l'investisseur ou de la plateforme de garde qu'il choisit.

À quelle fréquence les modèles sont-ils mis à jour ?

Les paramètres des modèles sont réévalués à intervalles réguliers en fonction des nouvelles données disponibles. Une revalidation par backtesting est effectuée avant chaque déploiement de mise à jour.

Comment le backtesting est-il conduit ?

Chaque modèle est testé sur des périodes historiques distinctes, incluant des phases de forte et de faible volatilité. Les résultats sont comparés à des scénarios de référence avant toute mise en production.

Virelo Xanum — visualization of market data streams analyzed by artificial intelligence

Artificial intelligence at the service of your digital assets.

Virelo Xanum aggregates continuous market data and applies predictive models to guide your crypto portfolio decisions, without emotional bias and with full signal traceability.

Access the platform Understand the methodology

A rigorous methodology, based on data.

Each recommendation produced by Virelo Xanum results from a three-step process, designed to limit subjectivity and document each decision.

Market data aggregation

Price feeds, volumes and inter-asset correlations are collected continuously from multiple sources, then normalized before any algorithmic processing.

AI predictive modeling

The models are trained on long histories and validated by successive backtesting, in order to assess their robustness over distinct market cycles.

Automated risk adjustment

A risk mitigation module recalibrates portfolio exposure based on observed volatility and measured correlations between assets.

Historical performance and predictive analytics.

The results presented come from backtests carried out on historical data. They illustrate past behavior of the models and are not a guarantee of future performance.

Evolution of signal success rate — backtest period

Data from internal backtests, sampled by comparable market periods.

Signal success rate Proportion of signals whose anticipated direction was verified over the backtest period.

Reduced drawdown Difference measured between the exposure adjusted by the algorithm and an equivalent static exposure.

Algorithmic stability Consistency of the model results over market sub-periods with different volatility regimes.

Detailed figures, calculation methodologies and reference periods are communicated on request, in a complete backtesting report.

Virelo Xanum — technical team analyzing risk models and market data

An approach built for prudent investors.

Virelo Xanum is aimed at investors who want to expose part of their portfolio to digital assets without giving up structured risk management.

The platform documents its hypotheses, publishes its calculation parameters and makes available the history of adjustments made by the models.

Functional tools to manage your exhibition.

Three functions cover the entire decision cycle, from signal detection to portfolio adjustment.

Real-time predictive analytics

Signals are recalculated with each market data update. You have an up-to-date reading without manual intervention.

Automated portfolio optimization

The allocation between assets is recalculated according to the measured correlations. The goal is to maintain a risk profile consistent with your parameters.

Dynamic risk management

Exposure is automatically reduced during phases of high volatility detected. This logic aims to limit the scale of potential losses.

Frequently asked questions about the method and safety.

The following answers specify the origin of the data, the frequency of model updates and the security conditions of the assets.

Where does the data used by the models come from?

The models rely on feeds of price, volume and liquidity indicators from multiple exchanges. These flows are aggregated and verified before integration into the predictive models.

How is asset security ensured?

Virelo Xanum produces risk allocation and adjustment recommendations, but does not directly hold user assets. Custody remains the responsibility of the investor or the custody platform they choose.

How often are the templates updated?

Model parameters are re-evaluated at regular intervals based on new available data. Revalidation by backtesting is carried out before each update deployment.

How is backtesting carried out?

Each model is tested over distinct historical periods, including phases of high and low volatility. The results are compared to reference scenarios before being put into production.