The third law of search engine
- Baidu President Li Yanhong Search Engine Today, it is already an end of the past and opens up the future. In order to explain the third law I told, let's take a look at the first and second laws. The law of the first fixed law sounds like an academic papers. It is indeed, even the first, second laws have never been previously, but the first, the second law is indeed since the industry and academia. Confirmation. In fact, this first definition is widely studied by the academic community before the Internet appears, that is, the so-called correlation law. At this field, the information retrieval, or information retrieval, and also called full text. At that time, the correlation was based on the word frequency statistics, that is, when the user enters the search term, the search engine goes to search the search term in the article (web page), the location is more important, plus Some weighted results for the degree of extent of the search term itself, and finally discharge a result (retrieve the results page). Early search engine results are based on the first diamonds of this article, such as Infoseek, Excite, Lycos, etc., which are basically the research results in the academic community before using the network era, and the main energy of the industry is dealing with large visits. On the large amount of data, there is no breakthrough in relevance. The word frequency statistics did not use any of the network-related characteristics, which is the technology of the former network age. However, the main literature in the network is existing in the form of a web page, and almost everyone can publish various contents online with the heart, the same number of words, the quality phase can be very far, but according to the first search engine The law, the sort of these two web pages should be the same. In order to be able to send some of the first few of the search results, the producer of many web content races the brain, stacked the keywords on its page, and the search engine will prevent this, and it is miserable. This situation has changed in 1996. The Law of the Second Law Personality Law In April 1996, I went to the gambling city of Las Vegas to open a academic conference on information retrieval. The content of the meeting is like the weather in Las Vegas, and it is more boring. But I am far from the company, but it is rare to have a chance to carefully think about the problem. Just when listening to a unhappy paper speech, I suddenly linked the mechanism of the scientific quotation index with the super link on the web - grateful to Peking University, she taught my scientific quarters when I was on my third year. Mechanism, the United States I am afraid that there is no university to teach this play in your undergraduate. The mechanism of the scientific quotation index, is white, who is a number of thesis, who is considered authority, the paper is a good passion. This idea is transplanted to the web, which is more than the number of links. The page is considered to be high quality and popular. In addition to the corresponding link text analysis, you can use it in the sort of the search results. This leads out the second law of search engines: the law of popularity. According to this law, the relevance of the search results is not completely dependent on the word frequency statistics, but more dependent on hyperlink analysis. I realized that this is a breakthrough, and I will soon summarize my idea after going back, and I applied for this aspect of U.S. Patent in June 96. On July 6, 1999, the US Patent and Trademark Office approved the patent number 5,920,859, in the patent of the only inventors. At the end of 1996, the two graduate students at Stanford University's computer system also thought of the same solution. They later found a search engine called Google, and the Google's website still said that their technology is patent-pending. In the application), I don't know if the US patent office will also batch such a patent. Anyway, the method of chain analysis is gradually accepted by major search engines after 98 years. Since the link is a fundamental feature of the network content, the search engine at this time began to truly utilize the search technology in the network age.
