Let novice friends better understand the simple hits algorithm

&http://www.aliyun.com/zixun/aggregation/37954.html ">nbsp; Today we will introduce the peak of the hyper-chain analysis: Hilltop algorithm, as now Google's core ranking algorithm is one of the online, there are a large number of literature introduced her. This article focuses on the analysis of the original algorithm, without considering too many complex factors, making it easier for you to understand the nature of the algorithm.

Hilltop algorithm set pagerank,hits, correlation algorithm in a large, by the Compaq System Research Center Krishna Bharat and the University of Toronto George A.mihaila in 2001 and applied for a patent, after authorization to Google, December 2003 Google algorithm update, which became one of Google's core ranking algorithm.

Search engine algorithm introduced hits algorithm. Hits algorithm is the most authoritative and widely used algorithm in web structure mining. The hits algorithm was proposed by Joen Kleinberg (Jon Kleinberg) in 1998, and the research of the algorithm inspired the birth of PageRank algorithm. The main idea of the hits algorithm is that the importance of Web pages is related to the subject of the query.

We can understand this: The hits algorithm is based on the topic to measure the importance of the page, relative to different topics, the same page is the importance of different degrees. For example, Baidu for the theme "Search engine" and the theme "Hunan SEO" the importance of different.

The hits algorithm uses two important concepts: the authoritative Web page (authority) and the Central Web page (hub).

For example: Google, Baidu, Yahoo!, Bing, Sogou, Soso and other search engines as opposed to the theme of "search Engine" is the authoritative page (authority), because these pages will be a large number of hyperlinks point.

Http://www.pyy1990.cn/post/Hits-Algorithm.html This page links to these authoritative pages (authority), this page can be called the theme "search Engine" Center page (hub).

The hits algorithm found that in many cases, there was no link between authoritative pages (authority) under the same topic. Therefore, authoritative Web pages (authority) are usually associated through a central Web page (hub).

The hits algorithm describes a dependency between an authoritative Web page (authority) and a central Web page (hub): A good central Web page (hub) should point to many good authoritative pages (authority), and a Good authoritative web page (authority) Should be pointed to by many good central Web pages (hubs).

Two problems that arise at the same time are:

The hits algorithm separates the link from the content, considers only the link structure between the pages to analyze the authority of a page and another page, for example, for navigation or for paid advertising.

The solution to the first problem is to use the hyperlink text and its surrounding text to match the keyword to compute the hyperlink weights, and introduce the relative control of the weights to the surrounding text and the hyperlink text by the coefficients.

The solution to the second problem is that the hits algorithm introduces a time parameter, that is, to evaluate whether it is a normal reference by the length of the reference to a link.

On the principle of hits algorithm, the previous article has been briefly introduced. In fact, the hits algorithm is quite complex, not a few words can be summed up. This paper is collected and sorted out, aiming to make the novice friends like Xiao Peng better understand the simple hits algorithm.

Hilltop is a query dependency link analysis algorithm, which overcomes the disadvantage of PageRank query independence. Simply put, the hilltop algorithm is an algorithm for reordering search results for popular query keywords. This is due to the low efficiency of the hilltop algorithm for popular keywords. The algorithm is divided into two main processes:

The search and grading of the expert pages; Search engine based on user query log found hot keywords, start to search for these hot keywords expert page, become an expert page of 2 necessary factors, 1 must have enough and no affiliation of the chain, 2, there is at least a phrase containing the hot keyword all terms. After you have identified the expert page, find all the terms in the list that contain the most popular keywords, or 1 to 22 terms of the phrase, the phrases are divided into three grades, respectively, all inclusive, the difference of 1 and 2 terms, respectively, for this three-level calculation of grade, grade is divided into each level of all the phrases score and, And the phrase score depends on the position of the phrase in the page, the score from high to low in order the title, head and anchor text, and so on, and then the comprehensive calculation of this three-grade score will be experts. Here's a simple example of "car consumption", the HOT keyword. China auto consumer Network "home page and Friendship link is the key word of the expert page, because he has enough and not subordinate to the 315che.com host domain name and the same as the C-class IP out of the chain, and the title of" China Auto Consumer Network "also contains The terms "car" and "consumption". Next, score the first rank (the phrase that contains all the terms), the phrase "China automobile consumer Net" obtains 16 points in the title (hypothesis), as well as in the anchor text "China automobile consumer finances the tendency Big Survey" to obtain 1 points, then the first grade must divide into 17 points, then calculates the second rank (to be inferior one term), Third level (two terms). In this way, the weighted sum of three grade points is the expert score.

Second, the target page score; an expert page's score on the target page equals the expert's own score x the number of phrases that can be distinguished by the expert page. Take the top n point to the target page of the expert page, for many of the same subordinate experts page to point to the target page, take the highest score of the expert page, and then these experts page on the target page score and get, this page corresponds to this popular keyword score, some people call industry score.

We can see that the hilltop algorithm ensures the relevance of the evaluation results to the key words through different grades, and ensures the relevance of the subject (industry) through the grading of different positions, and prevents the keyword from piling up by the number of distinguishable phrases.

Summary: Hilltop algorithm exists a game of thought, in the link of the same industry site needs both competition and cooperation, only by peer "recognized" the site of the hot key keyword queries will be ranked in the front. Hilltop basically destroyed the small web site on the popular words of hope, unless you have a strong ability to anticipate the hot keyword, but this flow will only last a very short time. Of course hilltop is only an important factor in the rankings, not all of them.

Original: Xiao Peng @ changsha seo http://www.pyy1990.cn/reprint please keep.