Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall
Google's PageRank is not much introduced, a measure of the importance of the page algorithm, is essentially the result of the Web page mutual vote, based on this feature, we can use Sitemap to allow search engines as much as possible to browse the content of the site, but also through the chain to improve the PR value of the site, To achieve the purpose of SEO.
Most of the search engines in the market are using PageRank similar methods, and in order to ensure fairness, are used purely machine-run way, through web crawler to traverse the Web site, which has some interesting problems:
1, a Web page content is very good, but because the chain is too small, the crawler in the set depth threshold may not be able to climb to it, become a few people's "dark content"
2, some of the site because the PR value is high, even if the content or value of the content is not high, there may be a good search rankings, even if the technology-leading search engine using the semantic network method to identify quality content, the effect is still not good enough
In order to avoid the above problems, the introduction of user data to judge the importance of Web content and quality, is a research direction, how to do it?
Hypothesis: Browsing behavior in a timely manner is the best judge of the quality of Web pages, the equivalent of user labeling, in large-scale data, the effect should be superior to the machine
Principle:
1, through the browser or other client software, the best firewall or other security software, access to user browsing logs, to the search engine crawler database, that is, to get users to browse the data
2, the crawler matching existing index library, to find the contents of the index, crawl
3, the use of user log to the Web page to vote, the longer the browsing time weight higher, calculate the page rank
Defects:
1. Dependent client
2, there are user privacy issues
Avoid:
1, put forward Cloud antivirus, cloud defense, cloud security, let users agree to upload browsing records
2, secretly upload, will browse records (other files can also) encrypted and split upload, in the server-side combination restore
Well, the idea is finished, to give it a loud and profound name: Peoplerank
Finally, I'm very serious about technology.
Via I black Horse by Sluke Lu Weizing original address: Http://luplusplus.com/peoplerank-modle