kshape

[guanxin1121]

some questions regarding your kshape code:

1. can your code be applied to time series with different lengths?
2. I need to extract shape prototype, but could not figure out what inputs are needed? Would you explain the parameters in _shape_extraction function?

[Mic92]

1. Yes. The time series must be of same length. You can try to interpolate values if you have missing intermediates.

# @idx the assignment of each timeserie to a specific cluster
# @x all timeseries
# @j index of the cluster that the shape is computed for
# @cur_center the current centroid on which the new centroid will be computed
def _extract_shape(idx, x, j, cur_center):
  pass

[johnpaparrizos]

Interpolation for missing values or padding of zeros left/right of the time series to bring them in the same size (this is what FFT is doing either way internally).