Trend Removal

Problem

Remove a known linear trend from a series: x[t] - (intercept + slope * t) for t = 0, ..., n-1.

Return the detrended series. Neural forecasters often model the detrended residual and add the trend back.

Hint

Subtracting the fitted line makes the series stationary in the mean.

case 1

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

expected: [1,1,1]

case 2

in: [[10,20,30],10,10]

expected: [0,0,0]

case 3

in: [[5,5],0,5]

expected: [0,0]

case 4

in: [[],1,1]

expected: []

Starter codePython

def trend_removal(series, slope, intercept):
    # Your code here
    pass

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