Direct and Recursive Blend

Problem

Combine a recursive multi-step forecast and a direct forecast: weight * direct_value + (1 - weight) * recursive_value.

Return the blended value. weight = 1 uses only the direct forecast.

Hint

Blending hedges the bias of recursive extrapolation against direct-model variance.

case 1

in: [10,12,0.5]

expected: 11

case 2

in: [10,12,1]

expected: 12

case 3

in: [10,12,0]

expected: 10

case 4

in: [3,5,0.25]

expected: 3.5

Starter codePython

def direct_recursive_blend(recursive_value, direct_value, weight):
    # Your code here
    pass

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