Introduction to the common algorithm model for SEO must know

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

A lot of webmaster friends publicize SEO really very simple, change code, do links, write articles on the line, one months can let novice learn. Let me talk about the views of these people, I think these people simply do not understand SEO, is a layman, just know a little bit of fur, purely by a little experience, do not know anything deep. SEO is what?—— search engine optimization, you are not familiar with the search engine, do not understand the search engine rankings algorithm, then you also optimize the fart ah.

Overseas do SEO is not the same as the domestic, foreign generally more in-depth, domestic seoer is generally linked to income, the pursuit of short-term benefits, a superficial understanding of search engines, according to experience. Of course, search engine algorithm is a very complicated learning, I dare say that no one in the world can really get through the algorithm, of course, except for search engine engineers, haha. Personally think that do SEO as long as in-depth understanding of search engine basic algorithm model on the line, I am here to put these models simple introduction, everyone if you want to discuss with me, welcome to add my QQ.

One, inverted calculation:

A collection of keywords for Web pages and documents after the inverted row is a huge collection. which can represent the page or the content of the document which keyword is the most weight of keywords.

Of course, search based on this there are a variety of algorithms can be used;

Second, the calculation based on Boolean model:

The real Boolean model is the numerical calculation of 0 and 1, which has high rate but accurate rate.

Three, based on the vector model of the calculation:

Using vectors, the correlation of each keyword to the vector is computed. The accuracy of customer inquiries has been greatly improved. The computation of TF is based on the normalized word frequency of relative word frequency, which is mainly the algorithm.

Calculation based on probability model:

Basic idea: Given a user's query string, there is a collection of all related documents relative to the string. We view such collections as an ideal set of result documents that can easily get results. This allows query processing to be treated as a set attribute of an ideal result document.

Five, based on Bayesian law calculation:

P (Dj|r) evaluating (R)

SIM (dj,q) =------------

- -

P (Dj|r) evaluating (R)

Vi. Generalized vector space models:

The index word vectors are linear independent rather than 22-intersecting. In the generalized vector model, two keyword vectors may not be orthogonal, which means that the index word vectors cannot be considered as orthogonal basis vectors of vector space, but are composed of smaller components instead.

Seven, neural network model:

Artificial neural network is an abstraction and simulation of some characteristics of human brain or natural neural network. It is mainly based on the results of physiological research in the brain, and aims to simulate certain mechanisms and mechanisms of the brain to achieve some function. The human neuron consists of three parts, the cell body, the dendrites and the axon.

Viii. Inference Network Model:

This is a new retrieval model based on Bayesian confidence network theory.

Nine, trust degree network model:

Another retrieval model based on Bayesian network theory is the Trust degree network model.