Robust Scaling

Scale each feature (column) by centering it at its median and dividing by its interquartile range $Q_3-Q_{1}Q\_3-Q\_1$ (quartiles use linear interpolation).

Implement solve(X).

Returns. Return a NumPy array of the same shape. A column whose interquartile range is zero maps to all zeros.

Examples

Example 1

Input:

solve([[1.0], [2.0], [3.0], [4.0]])

Output:

[[-1.0], [-0.333333], [0.333333], [1.0]]

Example 2

Input:

solve([[0.0], [0.0], [0.0], [9.0]])

Output:

[[0.0], [0.0], [0.0], [4.0]]