The development direction of the search engine

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

The value of Internet existence is low cost, high capacity, information transmission of various parties.

The internet every killer application is inseparable from the information and delivery of these two keywords. The mailbox is, the instant communication is, the search engine or, the future killer application still can not leave the information and transmits these two key words.

The development history of the search engine is a process of excavating the user's needs and satisfying the user's needs. In the foreseeable future, from the product point of view of the development of Web search engine has the following several aspects:

1. Accurate user's intention information extraction, optimization sorting

Users in the search for "the latest", "Free", "official website", "Beijing", "telephone" and other keywords when not necessarily need to have the Web page in this keyword, but to find such information.

Users looking for "the latest" actually want to get the latest relevant content of other words, not necessarily the need to contain the "latest" two words. So in the sort of time to consider the Web page arrangement in front of the position more satisfied with the user's needs.

User Search "18 Street twist Beijing" is to find in Beijing's 18 street twist address or phone.

Users search for "Beijing Ze Tong Hua Cheng Technology Development Co., Ltd. Phone" is looking for the phone number.

In processing this kind of request needs to the geographical information and "The telephone" this kind of vocabulary carries on the front-end analysis, in the index time identifies the telephone number, the address information, in the sort time will have the relevant information the page to place in the front, and when makes the summary extraction time directly manifests the user needs.

2. Block analysis based on Visual Web

This technique is exciting and helpful for optimizing the sorting of Web pages and the quality of automatic summaries. Web search engine can be full-text search as in almost pure data processing, plus the rich information on the Web page, you say that the relevance of web search can not be significantly improved?

3. Web page Library content classification

Users in the search for "Shenhua", that he may be two requirements, 1. Football related 2 Shenhua Electric 3. Other

If the user searches "Shenhua" to come out all is the football related information, this obviously cannot represent the different Netizen's demand. As an entrance, if the different types of information (industry different, different knowledge types) in the home page, the user will feel very happy, meet the needs of diversity.

This can also be a preparation for personalized search in the future.

4. Potential relevance

Search "Horror", appeared a bin Laden news, although this article has no "horror" this keyword.

Search for "tomatoes" appear "tomatoes", but there is no "tomato" in the Web page keyword.

This technique seems to be immature.

5. Web page Structured Information Extraction technology, Web page on the relevance of text content analysis

Structured information extraction is the best technology in the future, and automatically extracts structured data on any Web page. The main available vertical search engine: The Web page data collection, extraction, deep processing to provide users with better, more professional services.

Structured information extraction can identify the correlation between text in a Web page, and can be used to improve the relevance of multiple lexical retrieval (calculating offsets not only on text distances but also in table cells), improving the relevance of links, and improving the relevance of files and texts ...

Map search, Yellow Pages search, mp3 search, image search, BBS search and so on all kinds of search can not be separated from the Web page structured information extraction.

6. Natural language processing, simple semantic grammar analysis

NLP still has a long way to go, and in the course of walking you can get a lot of value to use. NLP might not have been very successful, but it could have spawned a new technology that was very successful.

NLP does not need to be completely successful before it can be used.

Search engines can make simple grammatical analysis based on the content and present some to the user. For example, Google's "DEFINE:" The use of this method, the identification of synonyms can be used to this simple grammar analysis to deal with! It can also make key words in the form of some kind of grammar, and improve the retrieval effect.

7. Duplicate recognition

The data redundancy of the Internet is too great, an article may be reproduced thousands of tens of thousands of times.

Identify duplicate Web sites, Web pages, duplicate body text, duplicate paragraph recognition ....

Let the user feel "Wow! The content here is not repeated!"

At the same time to the repetitive information to adjust the weight of the information is generally more popular, should have a higher weight value. But to the content of the news category to identify, within a certain period of time weighted, a certain amount of time to fall right.

8. Industry optimization

The industry of search engine is unavoidable. The only hurdle that affects the search engine industry is technology or difficulty (the technology here is not the sort of pediatric template based Meta Data Acquisition Word index).

But the web search engine can be the greatest degree of industry, in this point Baidu appears Zhuo far-sighted. Build Baidu know not only can enrich content, corpus, fasten user, even profit. More use can use Baidu know the professional Search user group of various industries to improve the search for various industries Baidu user analysis of the effect of each industry, it is true that the meaning of the users of the industry Baidu can be very low-cost access to mobilize professionals to optimize the effect of Baidu can do.

9. Related Search

A friend said to me a few days ago. The main function of "related search" is two, 1. Tips for users to search the words of other netizens (help not to select keywords, user choice of keywords, provide an interaction between users) 2. Recommend more relevant search terms with better results

The first function is basically satisfied. The second search engine is largely out of place. How to complete the second function is difficult. But to a certain extent, it's easy.

10. Gather more data

The data on the Internet is only a small part of the data in the world, and search engines are not satisfied with the speed at which the ants move their bricks. Through a low-cost and efficient data acquisition method to capture the information under the line, the human brain is the search engine companies chasing.

Spider Manufacturing + User manufacturing + manufacturing + co-manufacture

11. Tracking the changes in the Internet, the details of the optimization, game

Search engine is an application which is closely related to internet websites and netizens, and its data comprehensiveness is closely related to data source and collection system.

For the changes in the structure of the Web page, content changes, the needs of netizens change, needs to be continuously improved. All kinds of details of the improvement are search engine difficulties, but also must go the road, search engine development is to pay attention to details, a problem to solve.

And the game of popular search engine optimization. This article is supplied by HTTP://WWW.DIGCARS.COM.