vpin_proxy

from quantmaster.features.microstructure import vpin_proxy

df["vpin_20"] = vpin_proxy(df, window=20)

Proxy for VPIN (Volume-Synchronized Probability of Informed Trading).

Uses candle direction to classify volume as buy/sell. VPIN ~ |V_buy - V_sell| / (V_buy + V_sell) (rolling sum)

Source code in src/quantmaster/features/microstructure.py

def vpin_proxy(
    data: pd.DataFrame,
    *,
    window: int = 20,
    open_col: str = "open",
    close_col: str = "close",
    volume_col: str = "volume",
) -> pd.Series:
    """
    Proxy for VPIN (Volume-Synchronized Probability of Informed Trading).

    Uses candle direction to classify volume as buy/sell.
    VPIN ~ |V_buy - V_sell| / (V_buy + V_sell) (rolling sum)
    """
    window = validate_positive_int(window, name="window")
    validate_columns(data, required=(open_col, close_col, volume_col))

    o = pd.to_numeric(data[open_col], errors="coerce").astype(float)
    c = pd.to_numeric(data[close_col], errors="coerce").astype(float)
    v = pd.to_numeric(data[volume_col], errors="coerce").astype(float)

    signed_volume = v * np.sign(c - o)

    # Absolute imbalance sum
    num = signed_volume.rolling(window).sum().abs()
    # Total volume sum
    den = v.rolling(window).sum()

    out = num / den.where(den > 0)
    out.name = f"vpin_proxy_{window}"
    return out