Chapter 12. An approximate algorithm of definition of the document relevance to the inquiry.

It is necessary to hear the request to open "the magic formula" with which help it is possible to achieve good positions in search systems from the people who were recently engaged in search optimization. It is possible to give a question about density of keywords on a page, necessary for hit on the first lines of delivery. At once I wish to warn you - such formulas and magic figures are not present.

It is false that algorithms of search system are absolutely not available for «simple people». Yes, these algorithms will never be known completely because a great amount of people wishing to make such «ideal» pages and hit the top in searchers. It often decreases a level of search quality.

However, the main principles of algorithm working are known. Before we start to tell about every factor influencing on relevance of documents to enquiries, I want to tell you about common formula approximating ranking formulas used by the most popular search engines. As I said, it was not a concrete formula using by search engines, but an extended formula describing a process of relevance detection of a document to an enquiry. It is given below.

Rа(x)=(m*Tа(x)+p*Lа(x))* F(PRa),

Rа(x) - total conformity of the document a to inquiry x,
Tа(x) - relevance of the text (code) of the document a to inquiry x,
Lа(x) - relevance of the text of links from other documents on the document a to inquiry x,
PRа - indicator of authoritativeness of the page a, a constant accorfing to х,
F(PRa) - The function not decreasing monotonously, F (0) =1, it is possible to admit that F(PRa) = (1+q*PRа),
m, p, q – some coefficients.

Certainly, this formula gives too general idea about algorithms of documents ranging in results of search and even can make a question - "why if all searchers use similar algorithm, results in them frequently strongly differ?" As they say, "business in details". Any of these indicators is function from others which can be considered or not by search system and each of these indicators has "weight", i.e. factor, various for each concrete searcher. Also own resources of all catalogues of search systems can influence.

We will speak about it further, and now I wish to pay attention only to the total position of a site in results of search depends on 3 basic components:

  1. Code relevance with respect to an enquiry Tа(x);
  2. Relevance with respect to links inquiry on the pages from other pages (or linking ranking) Lа(x);
  3. The indicator of authoritativeness of page PRа - the factor not dependent on search inquiry. Also it influences on effect from both previous components.

We will start to consider more carefully the factors influencing documents relevance with respect to enquiries from this «indicator of authoritativeness»

2. Factors not depending on an enquiry (statistic)

This factor is called the indicator of authoritativeness or the range of a document. It is designated as PRa. It is named in the considered search engines differently. However, search engines uses algorithms taking into account hyperlinks between documents while they are calculating it. These algorithms are modifications of the algorithm PageRank.