Anomalous Volume Dynamics Analyzer

Instant changes in the level of trading volumes are significant and very accurate indicators of potential changes in the sector.

We conclude, on the basis of a large number of experiments, that a sudden rise in trading volumes is always followed by a large-scale activation of the market.

So we built an assistant that analyzes real-time trading volumes on the smallest available timeframes and detects each person increase in the amount of trading. This allows anomalous cases to be clearly documented, where the rise in volume greatly exceeds the normal trading result or a specific cryptocurrency.

For noise filtering, if the coin displays increased volatility for a long time, the algorithm smoothes all signals since, under such conditions, any deviations do not necessarily mean a shift in the mood of market traders.

The framework also analyzes important variations from the expected volume level as a special case. The expected volume is computed using the N-th degree polynomial regression, where N depends on the current situation and is automatically determined by the assistant.