codecademy_handwriting_recognition

Handwriting Recognition using K-Means

The U.S. Postal Service has been using machine learning and scanning technologies since 1999. Because its postal offices have to look at roughly half a billion pieces of mail every day, they have done extensive research and developed very efficient algorithms for reading and understanding addresses. And not only the post office:

• ATMs can recognize handwritten bank checks
• Evernote can recognize handwritten task lists
• Expensify can recognize handwritten receipts

But how do they do it?

In this project, you will be using K-means clustering (the algorithm behind this magic) and scikit-learn to cluster images of handwritten digits.