US 6,041,311 · Granted 2000-03-21

The Microsoft Patent That Taught Computers to Read Your Mind

Imagine a Netflix-like system that figures out what you'll like by studying people similar to you. This patent describes how a computer can automatically learn your taste by comparing your ratings to those of people with similar preferences, then use that data to suggest things you've never seen before.

Curated and verified by the IsItPatented Editorial team

The plain-English version

What it protects

The patent covers a system that stores user profiles (ratings and preferences) and item profiles in memory, calculates similarity scores between users, and uses those scores to predict what rating a user would give to an unseen item. Specifically, it protects the method of weighting neighboring users by their similarity factors and combining their ratings to forecast whether a target user will like a particular item. The claim also covers scenarios where items have features that get clustered and weighted to refine predictions.

Why it matters

This is a foundational patent for recommendation engines, the technology that powers Netflix suggestions, Amazon product recommendations, and countless other personalization features. By automating the process of finding like-minded users and inferring preferences, Microsoft secured intellectual property around a core mechanic of modern e-commerce and streaming. The patent was filed in 1997, before recommendation systems became ubiquitous, making it a pioneer in the space during the early web era.

Real-world use

Every time you see a Amazon recommendation saying 'Customers who bought this also bought that,' or when Netflix suggests your next binge-worthy show, you're seeing this patent's logic in action.

Original USPTO abstract

A method for recommending items to users using automated collaborative filtering stores profiles of users relating ratings to items in memory. Profiles of items may also be stored in memory, the item profiles associating users with the rating given to the item by that user or inferred for the user by the system The user profiles include additional information relating to the user or associated with the rating given to an item by the user. Similarity factors with respect to other users, and confidence factors associated with the similarity factors, are calculated for a user and these similarity factors, in connection with the confidence factors, are used to select a set of neighboring users. The neighboring users are weighted based on their respective similarity factors, and a rating for an item contained in the domain is predicted. In one embodiment, items in the domain have features. In this embodiment, the values for features can be clustered, and the similarity factors incorporate assigned feature weights and feature value cluster weights.

Patent details

Publication number: US 6,041,311
Filing date: 1997-01-28
Grant date: 2000-03-21
Assignee: Microsoft Corporation
Inventor(s): CHISLENKO; ALEXANDER, LASHKARI; YEZDEZARD Z., MCNULTY; JOHN E.
CPC class: H04N21/252

View original record on Google Patents →

Want to file your own patent?

If you're building a recommendation feature or personalization tool of your own, search our database to see what prior art exists in this space before you design your algorithm.

Free patentability scan