When doing search engine optimization (SEO), the relevance of links is often mentioned when building external links. After all, most of your external links are for search engines, and search engines rely on spider programs. Come and get it.
Remember, a "spider" is a crawling "program", not an "artificial intelligence". He will not use human judgment logic to judge correlation, but some "logic programs". After all, spider programs are not complicated (in comparison).
After understanding the above situation, let me talk about my own views and construction experience on "relevance". Please correct me if I am wrong.
The correlation between your web page and the content in your web page is judged by the spider program:
Assume that your webpage is a new webpage, it has been created less than 2 days ago, and it has just been submitted. The spider will crawl it based on the "crawl index" (or through other forms of links).
First, the spider comes to your web page through various links. For him, the new web page features are judged from the elements in the <title> tag in the <head> part of the web page code.
Suppose the spider comes to your webpage and grabs the <title> text in the webpage: My Animation Forum-New Animation-New Animation Ideas!
The spider will first crawl the text in <title> as a "keyword" as a reference, and then look for related "keywords" from the web page code.
You can rest assured that the spider can judge some JS codes and tags in the page. It will pay more attention to the "keywords" in the title part of the text and div block, that is, a series of <h1>-<h4>. This can also verify what everyone said, why the content in <h1>-<h4> has a higher weight than the text content in other <body>. Then the spider will set the most repeated words in <tilte> in the <body> part as "reference keywords" (of course, there is a repetition limit, and the search engine will set and embed it in the spider program).
Then the spider itself will also judge the "keyword expansion" based on the "reference keywords" (based on the analysis and comparison of common data keywords in the relevant categories of search engines and the keywords in the title). Based on the above, determine the general "keywords" and do Reference for the next page crawl.
In the next step, the spider will calculate (this simple data calculation function is still available) the number of all keywords appearing in the web page code as the base number. Then calculate the number of "keywords" in the <body> tag as the numerator (sweat, I can't think of any words, popular ones come first). Then once divided, you will get the keyword density of your page.
After the density comes out, the search engine itself will have a standard judgment (artificially designed parameters). There will be a level that is the most reasonable, average or the worst. This standard is changing all the time. In short, it is based on data over a period of time. Adjustment (just to let you come back to your mind every once in a while).
Okay, after the density is determined, compare the keyword differences on this page.
Based on the keywords in <tilte>, the reference of "expanded keywords" is integrated, and the "keywords" captured in the <body> section are compared to compare the differences in each text. The differences are also graded. , those with little difference become long-tail keywords, and those with big differences are discarded as search keywords for this page. This is the same as the "keyword density" judgment method. In this way, the long-tail words and search keywords of the page are determined.
When two results are obtained for the two items of density and page keyword difference, there will be a calculation formula for adding weights (set by the engine itself and adjusted regularly based on the data) to obtain a score for "relevance" ( An algorithm similar to Baidu index) to determine the keywords of the page and determine the degree of relevance between the page content and the keywords.
This is how the correlation between the keywords of the page and the content of the page is derived.
So, how to judge the relevance of external links? Okay, let’s talk about it and go to part 2
Judgment of external links and web page relevance:
One-way link: A link from another webpage to yours.
On his webpage, the link anchor text description of your website must be related to, or similar to, the keywords that he links to your page. The spider's judgment method is the same as in the first part, except that the initial "keyword" reference , becomes the link anchor text of your web page on his web page.
Having said that, how relevant is this web page to your website? How does this spider judge?
This link anchor text becomes a bridge and reference.
The relationship can be understood as this: the keywords of your web page VS (relevance judgment, set to A) → anchor text (the anchor text that links to your web page) ← (relevance judgment, set to B) his web page keywords.
PS: For the method of judging A and B, please refer to Part One: The correlation between your web page and the content in your web page, and the judgment in the spider program.
After A and B are obtained, compare them. If the difference is within a certain range, the spider will have a grade standard (artificially set). For example, if the difference between A and B is within 10%, the correlation is the highest. The weight of the two web pages increases and becomes a weight calculation result (it may only affect the quality of the web page, and the impact on the ranking is not clear, to be seen); 10 %-15%..........Wait, I won't write it.
Friendly links:
When it comes to anchor text on the two pages, the methods are similar.
Keywords of your web page VS (relevance judgment, set to A) → anchor text (anchor text on one of the web pages) ← (relevance judgment, set to B) keywords of his web page
Keywords of your web page VS (relevance judgment, set to C) → anchor text (anchor text on another web page) ← (relevance judgment, set to D) keywords of his web page
Comparing A and B, a weight calculation result is obtained: E;
Comparing C and D, a weight calculation result is obtained: F
E and F are finally compared to obtain the final weight calculation result.
Thus, it is judged that the two web pages are exchanging links and the relevance of the two web pages is judged.
The above situation is suitable for one-way link correlation judgment and friendly link correlation judgment.
My experience, if there are any mistakes, please feel free to correct me!