Skip to content

Kernelized mean change (CostRbf)#

ruptures.costs.costrbf.CostRbf #

Kernel cost function (rbf kernel).

gram property readonly #

Generate the gram matrix (lazy loading).

Only access this function after a .fit() (otherwise self.signal is not defined).

__init__(self, gamma=None) special #

Initialize the object.

Source code in ruptures/costs/costrbf.py
def __init__(self, gamma=None):
    """Initialize the object."""
    self.min_size = 2
    self.gamma = gamma
    self._gram = None

error(self, start, end) #

Return the approximation cost on the segment [start:end].

Parameters:

Name Type Description Default
start int

start of the segment

required
end int

end of the segment

required

Returns:

Type Description
float

segment cost

Exceptions:

Type Description
NotEnoughPoints

when the segment is too short (less than min_size samples).

Source code in ruptures/costs/costrbf.py
def error(self, start, end) -> float:
    """Return the approximation cost on the segment [start:end].

    Args:
        start (int): start of the segment
        end (int): end of the segment

    Returns:
        segment cost

    Raises:
        NotEnoughPoints: when the segment is too short (less than `min_size` samples).
    """
    if end - start < self.min_size:
        raise NotEnoughPoints
    sub_gram = self.gram[start:end, start:end]
    val = np.diagonal(sub_gram).sum()
    val -= sub_gram.sum() / (end - start)
    return val

fit(self, signal) #

Sets parameters of the instance.

Parameters:

Name Type Description Default
signal array

signal. Shape (n_samples,) or (n_samples, n_features)

required

Returns:

Type Description
CostRbf

self

Source code in ruptures/costs/costrbf.py
def fit(self, signal) -> "CostRbf":
    """Sets parameters of the instance.

    Args:
        signal (array): signal. Shape (n_samples,) or (n_samples, n_features)

    Returns:
        self
    """
    if signal.ndim == 1:
        self.signal = signal.reshape(-1, 1)
    else:
        self.signal = signal

    # If gamma is none, set it using the median heuristic.
    # This heuristic involves computing the gram matrix which is lazy loaded
    # so we simply access the `.gram` property
    if self.gamma is None:
        self.gram

    return self