Min-Max Normalization

Say one test is graded out of 50 and another out of 500. You cannot compare the raw scores fairly. Normalization fixes this by putting every score on the same 0-to-1 scale. Min-max normalization does exactly that. The smallest value in your data becomes 0. The largest becomes 1. Everything else lands proportionally in between.

1. 1Find the smallest (min) and largest (max) values in your data.
2. 2For each value, subtract the min, then divide by the range (max minus min).
3. 3The smallest value becomes exactly 0, and the largest becomes exactly 1. Everything else falls proportionally between them.

Use this before feeding numeric features into machine learning models that are sensitive to scale, like k-nearest neighbors or gradient descent. This way, a feature measured in the thousands (like salary) does not drown out one measured in single digits (like age).

• ⚠Do not normalize before splitting into training and test data. Doing so leaks information about the test set into training.
• ⚠Watch out for dividing by zero when every value in the data is identical (max equals min). That case needs special handling.