How Websites Get Ranked by Importance

This patent describes a computer method for scoring documents in a linked database, like the internet, by considering the importance of other documents that link to them, helping search engines find better results.

What it covers

This patent outlines a computer method for assigning an "importance score" to documents within a network, such as web pages (Claim 1). It works by first obtaining a collection of documents where some link to others. Then, it calculates a score for each linked document based on the scores of all the other documents that point to it (Claim 1). For example, if a highly-scored website links to your blog post, your blog post's score would increase. The patent also describes various ways to adjust these scores, such as considering how many links a linking document has (Claim 2) or even how likely a user is to access that linking document (Claim 3). This process helps in organizing and presenting documents based on their calculated importance.

What it doesn't cover

• —Does not cover ranking documents solely based on the words they contain, without considering incoming links.
• —Does not cover ranking documents where the score is not influenced by the scores of other documents linking to it.
• —Does not cover ranking systems that only count the number of links to a document, without adjusting for the importance of the linking document itself.
• —Does not cover ranking based purely on user behavior metrics like clicks, unless those metrics are used to adjust the linking document's weighting factor (Claim 7).
• —Does not cover ranking where a "random traversal" (Claim 10) does not assign rank based on how many times a document has been traversed.

The clever bit

The truly novel idea was to treat links as "votes" of importance, where a link from a more important page counts more than a link from a less important page. This recursive definition of importance, where a page's score depends on the scores of the pages linking to it, was a breakthrough for ranking information in vast, interconnected networks.

Why it matters

This patent describes the core principles behind PageRank, the algorithm that powered Google's initial success and revolutionized web search. Before this, search engines often ranked pages primarily by keyword matching, leading to easily manipulated and low-quality results. By using the link structure of the web to determine importance, this method provided a far more relevant and robust way to find information, fundamentally changing how people navigate the internet.

Real-world examples

• 1.Google Search (PageRank algorithm)
• 2.Academic citation indexing (e.g., impact factors)
• 3.Social network influence scoring
• 4.Recommendation systems for linked content