PELT Segmentation Cost

Problem

Compute the PELT objective for a segmentation: the sum of within-segment squared deviations from each segment mean plus penalty * (number of segments - 1). Changepoint indices mark segment starts; indices that are not strictly inside the series are ignored.

The cost of an empty segment is skipped.

Hint

The penalty term charges for each additional segment introduced by a changepoint.

Test cases

case 1

in: [[1,2,3,10,11,12],[3],0.5]

expected: 4.5

case 2

in: [[1,2,3,4],[],1]

expected: 5

case 3

in: [[5,5,5,5],[1,2,3],0]

expected: 0

case 4

in: [[1,9,1,9],[2],2]

expected: 66

case 5

in: [[1,2],[0,5],1]

expected: 0.5

Starter codePython

def pelt_segmentation_cost(series, changepoints, penalty):
    # Your code here
    pass

5 test cases · run your code against them in the editor