What at present individuation searchs is bit more insufficient

Overall and character, the Web individuation information in applying at searching engine recommends a service to be able to draw lessons from domain of traditional electron business affairs medium Web individuation information recommends a service. But, have and the search engine that the individuation in domain of contemporary electron business affairs recommends systematic likeness function gets rarely however the attention of people. That is to say, contemporary search engine still cannot offer individuation to search a result widely, the search of same to a of different user inquiry always is identical as a result, it and the user that refer inquiry have nothing to do. Accordingly, because retrieving level oversight the personalized demand feature of the user, be opposite so specific user, search engine often returns many hash. The main reason that creates this kind of appearance depends on contemporary search engine facing a few problems that recommend systematic likeness with traditional individuation information, specific as follows.

Above all, user demand is conveyed effectively hard. This basically has two reasons: On one hand, because be not professional user to lack the training that demand conveys commonly, cannot understand effectively so and convey oneself information requirement, the subjective understanding that the consequence of generation is an user often is not made clear. Figure ground says, this kind of phenomenon is the user cannot describe him to want what to look for, unless let him,see the thing that wants to search. On the other hand, this also comes from cannot get information of relevant user individuation correctly at the system, because the function that gets user individuation feature actively is not had normally in system of network Information Retrieval,this basically is, also did not ask the user must refer the restriction that personalized information ability uses at the same time. The final result of two afore-mentioned respects can produce a system cannot get the phenomenon of information of user individuation feature effectively.

Next, retrieve a result accurate change and retrieve change quickly between existence is contradictory. When the huge data that searchs engine in processing, the individuation of a lot of traditions recommends a technology to be able to produce serious function problem normally, algorithm and technology often lack these individuation in basically applying to website of traditional and small-sized business affairs to shrink goodly put a sex. Of course, people also offerred a few means of settlements, if dimension is decreased about, clustering and Bayes network, although can be solved on certain level,shrink the problem that put a sex, however these technologies often are the mode information in leaving line phase draw-out to give primitive data through be in, use these mode to get commendatory gather at online phase, so although these methods can reduce on-line to treat expenditure, but often arise recommend result inaccurate case, online calculative is complex at the same time degree also can increase as the grow in quantity of mode.

Finally, the retrieval of contemporary search engine intelligence changes a standard inferior still. Because search engine to cannot understand the semantics of webpage content truly, so pure favourable geographical position matchs to analyse a method with statistic with the term, inevitable understanding error is caused in the meeting on certain level. Should say, even if can use certain intelligence semantics to analyse a function, also expend as a result of what treat time and space breathtaking, in business so the index that use search props up a system in still cannot gain ground extensively. (Think of.cn of 100 million Europe)