New Google PageRank Ranking Algorithm 2

Original: Dan Thies
Compile: Karen
Part 2: explore the new Google PageRank algorithm
2-1. Theme trend of Google
2-1-1. Reasons for problems with the page level (PageRank) and Google's old algorithms
The idea behind page-level (PageRank) computing systems is to tell you which sites are most important through a "random motion" over the Internet. The system simulates the process of a random surfers following up and clicking a random link on a page, and pressing the "return" button to reach the deepest page. The higher the page level, the higher the chance that random network surfers will find it.

This idea is actually quite creative. The more external links a web page has, the more chance a web browser will find it. At the same time, in the page-level algorithm system, the more popular the page is, the more the imported link will benefit from the link-this is because any network surfers will find the opportunity to find these links.

In terms of research papers in specific fields, the page-level system is almost impeccable. For example, if a user queries a clustering set of papers (or web pages) on the study of molecular particle physics, a page-level algorithm will soon tell you how to query a given condition, which papers are most relevant to the specific query conditions and the most important ones? The reason is that these papers are cited more times than other papers.

If resources on the Internet have the same theme, this kind of work is perfect. But as we know, resources on the Internet cover millions or even more theme, and in people's real life, the query user is often looking for information with a specific topic. Therefore, although the page-level system considers all links, it ignores the topic nature of linked pages.

Google has tried to include the text content of the link into the ranking algorithm to overcome this limitation. However, a savvy search engine marketer fool Google's ranking algorithm by establishing keyword-filling links on the Internet. A new industry also came into being with PageRank-that is, paid exchange and transaction links from higher "page-level" pages.

If a website can purchase or import links from unrelated sites to improve its ranking, the page-level technology will no longer be able to provide high-quality search results for the vast majority of query conditions. We have reason to believe that when Google, the world's top search engine, finds that the quality of its search results is getting worse, it will not sit back.

2-1-2. New technology debut: Topic-Sensitive PageRank)