Feature Scaling Function | Dataford Interview Questions

Your question is Feature Scaling Function. Start with the requirements on the right.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Smartsheet machine learning pipelines may contain features with very different numeric ranges. Implement a function that transforms a feature matrix using either min-max normalization or population z-score standardization.

Formal Specification

Given X, a non-empty rectangular matrix represented as a list of rows, where each row is one sample and each column is one feature, and a string method:

For method = "minmax", scale each column to the range [0, 1] using (x - minimum) / (maximum - minimum).
For method = "standard", transform each column using (x - mean) / standard_deviation, where the mean and standard deviation are computed across all rows. Use population standard deviation, dividing variance by the number of rows.
If a column has no variation, output 0.0 for every value in that column.
Return a new matrix. Do not modify X.

Constraints

1 <= len(X) <= 10^4
1 <= len(X[0]) <= 100
X is rectangular
Every feature value is a finite number
method is either "minmax" or "standard"

Function Signature

def scale_features(X, method):