INNER CODE UNIT · Python
y
matrix-profile-foundation/matrixprofile · matrixprofile/core.py:416
y = np.append(np.flipud(query), np.zeros([1, n - m]))
y = np.fft.fft(y)
z = np.fft.ifft(x * y)
return np.real(z[m - 1:n])
def sliding_dot_product(ts, query):
"""
Computes the sliding dot product for query over the time series using
convolution. Note that the result is trimmed due to the computations
being invalid; the len(query) to len(ts) is kept.
Parameters
----------
ts : array_like
The time series.
query : array_like