Skip to main content
Volume profile factors analyze the relationship between price movements and trading volume to identify periods of unusual market activity and price-volume correlations.

​
Volume Profile Components

ComponentDescriptionCalculation
Volume TrendShort vs. long-term volume ratio(short_avg / long_avg) - 1
Price-Volume CorrelationCorrelation between volume and absolute returnsnancorr_1d_nb(volume, abs(returns))
Volume SurgeLogarithmic volume increase detectionlog1p(ratio) * surge_scale

​
Volume Profile API Classes

ClassDescriptionKey Features
RollingVolumeProfileFactorRolling window analysisDynamic window sizing, trend detection
VolumeProfileFactorCVCross-validated volume analysisWalk-forward validation, parameter stability

​
RollingVolumeProfileFactor

RollingVolumeProfileFactor(
    window: int = 365,
    minp: int = None,
    short_term_window: int = 30,
    surge_scale: float = 5.0,
    score_scale: float = 2.0,
)
Volume Profile variation factor over rolling window.
  • Volume Profile Variation : This method calculates volume profile variation by combining volume-based signals with price movement (correlation between volume and price returns) over configurable time windows. It highlights volume surges and volume-price trends to generate an indicator.
  • Signal Calculation : The signals are scaled and combined to give a final output by iterating over the historical data (closing prices and volume) using a rolling window.
Workflow:
  1. Calculate Volume and Trend : calculates the average volume over two time periods: a short-term window and the entire window. It also computes the volume trend, defined as the ratio of short-term to long-term average volume.
  2. Volume-Price Correlation : Calculates the correlation between volume — smoothed over the short-term window — and price returns over the same period. This helps determine if changes in volume are associated with price movement.
  3. Volume Surge Detection : Detect volume surges by comparing the short-term volume with long-term volume. This surge is amplified by the surge_scale factor.
  4. Signal Combination : Combine the volume trend, price-volume correlation, and volume surge to create a composite signal.
  5. Iterating Over Symbols : Applies the volume and price trend calculations for each symbol.
  6. Final Signal Calculation : Slice the closing prices and volume data to the most recent volume_window and computes the signals. The resulting signals are aggregated and scaled. Finally, the signal is passed through a hyperbolic tangent (np.tanh) for smoothing and amplified before returning the final value.
Method generated by attrs for class RollingVolumeProfileFactor.

​
Ancestors

  • systematica.models.base.BaseStatArb
  • abc.ABC

​
Instance variables

  • minp: int: Mininim period. Defaults to None, which means no minimum period is applied.
  • score_scale: float: Score scale coefficient. Defaults to 2.0. This factor scales the final score output.
  • short_term_window: int: Short term window. Defaults to 30.
  • surge_scale: float: Volume surge scale coefficient. Defaults to 5.0. This factor amplifies the volume surge detection.
  • window: int: The size of the rolling window. Defaults to 365.

​
VolumeProfileFactorCV

VolumeProfileFactorCV(
    training_window: int = 365,
    testing_window: int = 60,
    splitter: str = 'from_custom_rolling',
    custom_splitter: str | None = None,
    custom_splitter_kwargs: dict | None = None,
    short_term_window: int = 30,
    surge_scale: float = 5.0,
    score_scale: float = 2.0,
)
Volume Profile variation factor over cross-validation splits.
  • Volume Profile Variation : This method calculates volume profile variation by combining volume-based signals with price movement (correlation between volume and price returns) over configurable time windows. It highlights volume surges and volume-price trends to generate an indicator.
  • Signal Calculation : The signals are scaled and combined to give a final output by iterating over the historical data (closing prices and volume) using a cross-validation approach.
Workflow:
  1. Calculate Volume and Trend : calculates the average volume over two time periods: a short-term window and the entire window. It also computes the volume trend, defined as the ratio of short-term to long-term average volume.
  2. Volume-Price Correlation : Calculates the correlation between volume — smoothed over the short-term window — and price returns over the same period. This helps determine if changes in volume are associated with price movement.
  3. Volume Surge Detection : Detect volume surges by comparing the short-term volume with long-term volume. This surge is amplified by the surge_scale factor.
  4. Signal Combination : Combine the volume trend, price-volume correlation, and volume surge to create a composite signal.
  5. Iterating Over Symbols : Applies the volume and price trend calculations for each symbol.
  6. Final Signal Calculation : Slice the closing prices and volume data to the most recent volume_window and computes the signals. The resulting signals are aggregated and scaled. Finally, the signal is passed through a hyperbolic tangent (np.tanh) for smoothing and amplified before returning the final value.
Method generated by attrs for class VolumeProfileFactorCV.

​
Ancestors

  • systematica.models.base.BaseStatArb
  • abc.ABC

​
Instance variables

  • custom_splitter: str | None: Custom splitter to use for data partitioning. Defaults to None. If set, it should be a string that matches a custom splitter function.
  • custom_splitter_kwargs: dict | None: Additional keyword arguments for the custom splitter. Defaults to None. If custom_splitter is set, this should contain any necessary parameters for the custom splitter function.
  • score_scale: float: Score scale coefficient. Defaults to 2.0. This factor scales the final score output.
  • short_term_window: int: Short term window. Defaults to 30. This is the window size used for short-term volume calculations.
  • splitter: str: Default splitter to be used if custom_splitter is not passed. Choices are from_rolling, from_expanding, from_custom_rolling, from_custom_expanding. Defaults to “from_custom_rolling”.
  • surge_scale: float: Volume surge scale coefficient. Defaults to 5.0. This factor amplifies the volume surge detection.
  • testing_window: int: The size of the testing window. Defaults to 60. This is the window size used for testing the model.
  • training_window: int: The size of the training window. Defaults to 365. This is the window size used for training the model.