Google last updated, many webmasters found themselves in the search results pages significant drop in the rankings, many previously appeared in the first page to the website now fall into three, four or even ten away, the discussion will deal with this in a number of SEO forums very intense.
In fact, our own websites have appeared similar phenomenon in the inquiry website in Google results page supplement the number of methods in an article mentioned, IT technology Comments significant decline in the rankings from Google’s lost a lot of traffic. And last month, although it may be our server failures, such as not normal googlebot Index website, which have led some pages were Add additional results, or Google lowered the trust Comments IT index, but only a few days of server failures – it is not even completely meaningless stands, and only intermittent restart IIS – unexpectedly triggered so much volatility? Still aroused the suspicion of people, especially in light of the same server in the Exploration and Vista SEO world has not shown such a remarkable trend.
In many people’s analysis, that this may be due in Google’s ranking algorithm recent adjustments, the final page will be ranked on the original minimal impact on the right of Latent Semantic Indexing the major big improvement from my personal point of view, it seems to have a certain truth.
Then, how should the understanding Latent Semantic Indexing? Latent Semantic Indexing what impact site / web page in Google’s ranking?
What is Latent Semantic Indexing (indexing, LSI)?
As we all know, traditional search engine algorithm based on keywords, – of course, the key word here said not only the key word in that website also includes links Anchor Text of Keywords. — Through the search engine website Keywords Statistics specific location, density and link keywords in Anchor Text or URL keywords, which are given in accordance with the degree of matching with the users of the relevant search results, this is the next level of development before the computer search engine “to provide the necessary user as “the closest simulation.
However, we must see that from the perspective of natural language, even if we can not say that all the words (or phrases), but at least most of the term with words more than just features, at the same time, natural language also has more than one meaning of the word characteristics, so that search results and users will value the ideal of a lot of distance : Polysemy will lead the search based on accurate matching algorithm bear contained in the results of many users really want to find is not the content, and more than just a word used in the search based on accurate matching algorithm is omitted in the outcome of many users really want to find content.
This drawback of the traditional algorithm search engine could not be overcome, after all, machines algorithm is dead, as played by the Terminator governor Oxfam, as people do not understand, as analyzed. Although many search engines are now conducting the study of natural language, but Objectively speaking, the natural language understanding still remains at the research stage, from the real point to the use of trusted far short of the goal, or even a change in perspective, even if the development of natural language understanding to practical stage, whether its efficiency protection can also unknown, and the search engine, the efficiency of the factors must be taken into account is, they do not want to put in more servers and other network resources to complete the same task, but if users need to conduct a search, such as a half day to see the search results no different from search primer Qing nightmare. — Of course, it is only against??Google, the search engine may be great homemade “Chinese word segmentation and Technology” has been able to successfully resolve this issue.
Latent Semantic Indexing (LSI: indexing, LSI) is the search engine trying to weaken this, as far as possible drawbacks one of the feasible solutions. Latent Semantic Indexing through bypassing natural language understanding, a large number of samples taken statistical analysis to find different words (words, phrases) the relevance of search results further to enable users really want to look closer to the content, but also to guarantee the efficiency of search.
Latent Semantic Indexing (indexing, LSI) mechanism
Simply Theoretically, Latent semantic indexing (indexing, LSI) Implementation Mechanism is not complicated, it is nothing but the website included in the normal course of the index, has added a step: not just simple statistics, analysis website and link keywords, the website will be indexed in the database and Key words include the same in other parts of the same keywords or web pages than the right, to determine the different semantic web and web correlation between specific Keywords relevance, at the same time, will be of high relevance of the Semantic Web than on the analysis conducted to find a specific page in the relevant keywords, identify specific Web page that does not exist but despite its content related keywords. — Comparison of this passage is a mouthful, I do not know whether the statements clear.
Of course, the concrete implementation of the algorithm is much more complex. Needs attention yes, right Latent Semantic Indexing, probably most people’s attention is one of Semantic (Semantic related), but Latent Semantic Indexing of technical documentation are often more stressed Latent (potential, implied), and not simply related to the semantic significance. For instance, the “water”, with its semantics associated likely to be “hot” and “cold water” and the likes, but can be potentially relevant is the “steam”, “ice”, and so on, there are great differences.
How specific search engine completed a similar calculation is not our concern, but it is clear that, although the search engine itself does not know what exactly represent a word, I do not know what is the meaning of a word, but by Latent Semantic Indexing algorithm, and compared simply matching keywords, search engines can a more accurate way of judging a specific website content and the search of relevance, which gives users to find the content, or even From a certain point of view, closer to “person” analysis, look at the contents of the judgment methods.
Google algorithm in Latent Semantic Indexing (indexing, LSI)
In general, Google will be the first Latent Semantic Indexing (indexing, LSI) for its AdWords / AdSense advertising system for the analysis of specific Web pages and the relationship between advertising, which will be the most relevant text of the advertisement devoted to a specific website. Only then will be introduced to its ranking algorithm, however, for the first time, its weight is very small, in the last year of Jagger and BigDaddy update gradually upgrade its weight.
The reason, perhaps because the initial Latent Semantic Indexing (indexing, LSI) algorithm is not perfect, of course, can be used if Adsense rankings will be used for problems with the algorithm and the continuous improvement in the ranking algorithm to gradually increase its impact. After all, Latent Semantic Indexing only a computer simulation, not real sense of natural language analysis, and its search engine results still fall short of the goal.
Need to re-emphasize that, Latent Semantic Indexing (indexing, LSI) is the current search engine ranking algorithm Keywords supplementary matching technology from the perspective of optimism it can be said that it is very meaningful complement, but not replace the existing keyword matching algorithm.
Latent Semantic Indexing (indexing, LSI) how they affect rankings?
This requires from the point of view of two, first of all, the introduction of Latent semantic indexing (indexing, LSI) and increase its weight will be lower part of the web site rankings in the SERP, particularly those involving specific Key words but not its “semantic related” or “potential” Keywords’s website, so , a good optimizing before, such as keyword density with the best, in words matching algorithm simply ranked the leading Web search results pages may be the first few pages completely disappeared.
Secondly, there are the same impact on Anchor Text links, if a website / homepage RP-link most of the Anchor Text use the same keywords, links its value will be substantially diminished.
How to deal with Latent Semantic Indexing (indexing, LSI)?
It should be said that comments from the IT technical look at the specific situation, particularly its exploration and SEO compared to recent personal feeling attributed to the decline in ranking Latent Semantic Indexing (indexing, LSI) the impact of raising the weights, or a certain reason. — In this specific analysis (not quoted);.
So, how should deal with Latent Semantic Indexing (indexing, LSI)?
Honestly, is given here, only arrived on the basis of the above analysis, or speculation, the feasibility to be verified.
First, we must ensure that the ranking of websites, the need for a series of selected core Keywords neighboring Key words, as many enterprises registered domain name (or trademarks), in addition to the use of the domain name (or trademarks), but also the protection of a group of domain name registration (or trademarks), to more effectively protect the core Keywords At the same time, neighboring Key words will be properly Add corresponding website, or the establishment of these peripheral Keywords corresponding content. Of course, these peripheral Keywords be associated with the semantic core Keywords or a certain extent, linked.
Secondly, the appropriate inverse transform the Anchor Text links not only to let the core Keywords appear in Anchor Text, must also ensure that the surrounding Keywords in Anchor Text of the frequency. RP-link Key words that use all the core web site name or keyword approach is not advisable.