CN102693223A

CN102693223A - Search method

Info

Publication number: CN102693223A
Application number: CN2011100670796A
Authority: CN
Inventors: 潘燕辉
Original assignee: Individual
Current assignee: Individual
Priority date: 2011-03-21
Filing date: 2011-03-21
Publication date: 2012-09-26

Abstract

The invention discloses a search method comprising the following steps of: searching contents of interest submitted by a user in a client and based on the contents, generating a mark label corresponding to the contents; uploading an ID of the search user, the contents submitted in step A1 and the mark label to a server to be stored in the historical search database; obtaining a historical label and the corresponding user ID, computing matching degrees of the contents submitted in step A1 with contents under the historical label, sorting the contents based on the matching degrees and preferentially recommending other contents which have relatively high matching degrees under the historical label to the search user; and preferentially recommending other contents the total amount of which is small under the historical label if the matching degrees are the same.

Description

A kind of searching method

Technical field

The present invention relates to field of computer technology, relate in particular to a kind of information search method of high hit rate.

Background technology

Traditional search engines has two types basically:

1, recommend to the ratio of certain selection according to all users, when Search Results occurring, select many results preferentially to recommend.Certainly the algorithm that wherein also has other references, such as result's confidence level, the behavior judgement after the user selects etc.Because this search is based on most of people's selection, so hit rate is not high.

2, the situation search is an intelligentized search of taking all factors into consideration user context, hobby and environment, through the deep understanding to user view, under the various scenes of user's internet usage service, offers the properest search service of user.The situation search comprises 7 key elements (6W&1H); It stresses " is basis with people (Who) "; Customer-centric just; Time (When), place (Where), input (What), demand (Want), custom (How), background factors such as (Why) according to its search behavior calculate optimal Search Results by situation, this result are directly appeared through user's search situation again.But this search necessarily just can come into force based on pre-designed criteria for classification, pays close attention to A Fanda such as the user, and search engine must know that A Fanda is portion's film, and is portion's science fiction film, and the 3D your writing just might be recommended similar factor.

Summary of the invention

Technical matters to be solved by this invention is that the deficiency that is directed against prior art provides a kind of searching method---be people having the same habits search for; So-called people having the same habits' search; Essence is exactly to find and the identical or close people of search subscriber hobby; The thing that they like is preferentially recommended search subscriber, rather than classification is traditionally recommended.

The present invention adopts following technical scheme:

A kind of searching method comprises the steps:

A1, search subscriber is submitted own interested content in client, generates the markup tags corresponding with said content according to said content;

A2 uploads onto the server the ID of search subscriber, content and the said markup tags that steps A 1 is submitted to, is stored in the historical search database;

A3 judges in said historical search database, whether have a history tab, and the content under this history tab is identical or close with the content that steps A 1 is submitted to; If deny, then execution in step A4; If yes, execution in step A5 then;

A4; In said historical search database, whether have a historical content, the content of being submitted to steps A 1 is identical or close, if yes; Then obtain the ID at this historical content place and the other guide that ID is write down down thereof, recommend the other guide that is write down under this ID to said search subscriber; If, then do not recommend general Search Results to said search subscriber;

A5; The described history tab of obtaining step A3, and corresponding ID, the matching degree of the content under content that calculation procedure A1 submitted to and the described history tab; Sort according to matching degree, preferentially recommend the other guide under the higher history tab of matching degree to said search subscriber; Under the identical match degree, the other guide under the history tab that preferentially the content recommendation total amount is few.

Described searching method, operation below said steps A 2 concrete the execution:

Whether A21 has the ID of search subscriber in the historical search database of server, if yes, content and the markup tags thereof then steps A 1 submitted to are stored in the historical search database, if not, and execution in step A22 then;

A22, whether the inquiry search subscriber registers ID, if yes, then carries out after ID registers search subscriber as and set up ID, and the content of then steps A 1 being submitted to is stored in the historical search database together with the ID and the markup tags thereof of said foundation; If not, system is that search subscriber provides calculating ID, and content, said calculating ID and markup tags thereof that steps A 1 is submitted to are stored in the historical search database.

Said searching method, said client are computing machine or intelligent mobile phone terminal or palm PC.

Utilize " content ", " label ", " ID " like this, search subscriber is easy just have been found and its user with identical or akin pursuit, shoots straight, and can be applied to all spectra, and prior criteria for classification needn't be arranged.It is that a kind of elder generation goes to mate the people who has same hobby according to user preferences, and the selection that will have the people of same hobby is then shared to the user who has same hobby.And needn't worry whether this hobby is the known classification of system.

Description of drawings

Fig. 1 is a searching method process flow diagram of the present invention;

Fig. 2 sets up the process flow diagram of historical search database at server end for the present invention.

Embodiment

Below in conjunction with specific embodiment, the present invention is elaborated.

Be illustrated in figure 1 as the process flow diagram of people having the same habits' searching method of the present invention, comprise the steps:

A1, search subscriber A submits own interested content in client, and client generates a markup tags corresponding with said content according to its content of submitting to, and this markup tags can generate arbitrarily, only plays the mark effect; For example the content of user A submission is books B, books F, books T, reference table 1;

A2 uploads onto the server the ID of search subscriber A, content (books B, books F, books T) and the said markup tags that steps A 1 is submitted to, is stored in the historical search database;

A3 judges in said historical search database, whether there are so one or more history tab, and the content under this history tab is identical or close with the content that steps A 1 is submitted to; If deny, then execution in step A4; If yes, execution in step A5 then;

A4; In said historical search database, whether have a historical content, the content of being submitted to steps A 1 is identical or close, if yes; Then obtain the ID at this historical content place and the other guide that ID is write down down thereof, recommend the other guide that is write down under this ID to said search subscriber; For example in table 1, all there are not content books S, books G, books K one by one under any label; But existence one does not have the item of label under user " ID-Lao Pan "; Record books S, books G, books K under this, other books L, books H, books D under then will this item recommend the user.For not, then recommend general Search Results to said search subscriber, promptly return the Search Results of routine search engine;

A5; The described history tab of obtaining step A3, and corresponding ID, the matching degree of the content under content that calculation procedure A1 submitted to and the described history tab; Sort according to matching degree, preferentially recommend the other guide under the higher history tab of matching degree to said search subscriber; Under the identical match degree; Other guide under the history tab that preferentially the content recommendation total amount is few; Having submitted books B, books F, books T to such as search subscriber, two labels arranged in table 1---label ikjlojdsa and label 1238989 have comprised books B, books F, books T respectively fully, then preferentially recommend the other guide under these two labels; For example books J, books L, books Z; But because 1238989 times contents of label are less, so preferentially recommend books Z, next recommends books J and books L.And a label---label mhiuwergf is arranged, and only comprised books B, books T, then its matching degree is lower, as less important recommendation items.

Table 1 historical search database storing structure

The present invention is based on user's input, search the people of same hobby earlier, and then come return results according to selection or information that the people of same hobby has.When setting up search database, the general search engine all is to utilize some programs data collection on the internet, and people having the same habits' search notes with user being that unit sets up database more except this collection mode.Such as searching music; Will be under user's agreement and the situation of knowing the inside story, each user is set up a unique ID on server (be divided into registered user and nonregistered user, nonregistered user will be calculated unique code automatically according to user computer; Automatically the user is classified); Simultaneously, under user's informed consent situation, the music information that it had uploaded onto the server to be stored in the historical search database.When the user used people having the same habits to search for, the music information that the label that also can be set up when at every turn using and label relate to uploaded in this user's data storehouse equally.

People having the same habits' search application scope of the present invention is extremely wide, and prerequisite is to collect the behavior of ability representative of consumer hobby, such as having downloaded some resource; Such as the webpage of often visiting; Such as the article of often buying, such as some thing of collection, such as crowd who often exchanges or the like

Equally, people having the same habits' search and sight search can have good fusion, such as in the time can not navigating to detailed programs, can carry out large-scale people having the same habits' search according to existing classification, can obtain good result equally.Such as often visiting certain on-line shop as the user, though do not buy concrete article, can recommend him and equally often patronize other on-line shops that the people of this on-line shop often goes, the article of perhaps buying still have good hit rate

Should be understood that, concerning those of ordinary skills, can improve or conversion, and all these improvement and conversion all should belong to the protection domain of accompanying claims of the present invention according to above-mentioned explanation.

Claims

1. a searching method is characterized in that, comprises the steps:

2. searching method according to claim 1 is characterized in that, operation below said steps A 2 concrete the execution:

3. according to the said searching method of claim 1, it is characterized in that said client is computing machine or intelligent mobile phone terminal or palm PC.