CN102929978B

CN102929978B - Based on the drop-down prompt system of input prefix

Info

Publication number: CN102929978B
Application number: CN201210395091.4A
Authority: CN
Inventors: 常富洋; 秦吉胜
Original assignee: Beijing Qihoo Technology Co Ltd; Qizhi Software Beijing Co Ltd
Current assignee: Beijing Qizhi Business Consulting Co ltd; Beijing Qihoo Technology Co Ltd
Priority date: 2012-10-17
Filing date: 2012-10-17
Publication date: 2015-07-29
Anticipated expiration: 2032-10-17
Also published as: CN102929978A

Abstract

The invention discloses a kind of drop-down prompt system based on input prefix, it comprises client, at least one memory node and agent node; Client is suitable for receiving search target keywords request and the request of described search target keywords being sent to agent node, comprises current input prefix in the request of search target keywords; Agent node comprises inquiry forwarding module, is suitable for described search target keywords request forward to corresponding memory node; Data forwarding module, the candidate item of the current input prefix of the correspondence being suitable for memory node to return is forwarded to client, and the candidate item of corresponding current input prefix is the drop-down prompting result of recommending probability large; Memory node is suitable for based on the request of described search target keywords, obtain and the candidate item sending corresponding current input prefix to agent node.This system can make drop-down prompting result arranged in sequence more can meet the search intention of user, improves accuracy.

Description

Based on the drop-down prompt system of input prefix

Technical field

The present invention relates to search technique field, be specifically related to a kind of drop-down prompt system based on input prefix.

Background technology

Along with the development of Internet technology, the information on internet presents volatile rising tendency, causes user must go to obtain by the mode of search the information of needs.Search is numerous netizens' obtaining information, and the main channel of access destination website.

Existing way of search adopts the rule of prefix matching, to the Search Results meeting prefix, is prompted to user according to drop-down after page browsing amount or click volume (Page View, PV) sequence from more to less.PV normally weighs the leading indicator of an Internet news channel or website or even a concerned degree of Internet news.As shown in Figure 1, when user's input " h ", drop-down prompting result sorts from more to less according to PV." hold live like " makes number one, and represents maximum with the searching times of " hold lives to like " in " h " Search Results that is prefix; " hotmail " comes the 5th, represents to come the 5th with the searching times of " hotmail " in " h " Search Results that is prefix.If inputted " o " after " h ", putting in order of " hold living to like " and " hotmail " is constant again, still for " hold lives to like " comes before " hotmail ".

Existing way of search is according to the overall PV statistical information of search engine, by drop-down prompting according to PV order arrangement from more to less, if do not meet the keyword of user search intent in drop-down prompting, although or in the keyword of drop-down prompting, there is the keyword can expressing user search intent, but this keyword comes position more on the lower in drop-down prompting, range search frame is far away, user can not be met the keyword of search intention according to the input prefix of input, although or there is the keyword meeting search intention, but also need user find in drop-down prompting and select this keyword, the accuracy that the drop-down prompting inputting prefix based on user meets user search intent is low.

Summary of the invention

In view of the above problems, the present invention is proposed to provide a kind of drop-down prompt system based on input prefix overcoming the problems referred to above or solve the problem at least in part.

According to an aspect of the present invention, provide a kind of drop-down prompt system based on input prefix, it comprises client, at least one memory node and agent node:

Wherein, client, is suitable for receiving the request of search target keywords, and the request of search target keywords is sent to agent node, comprise current input prefix in the request of search target keywords;

Agent node, comprising:

Inquiry forwarding module, is suitable for search target keywords request forward to corresponding memory node;

Data forwarding module, the candidate item of the current input prefix of the correspondence being suitable for memory node to return is forwarded to client, and the candidate item of corresponding current input prefix is the drop-down prompting result of recommending probability large;

Memory node, is suitable for based on search target keywords request, obtain and the candidate item sending corresponding current input prefix to agent node.

According to a kind of drop-down prompt system based on input prefix of the present invention, be in the input prefix branch of root in fixed current input prefix, according to the occurrence number of each input prefix, statistics obtains the weight of each input prefix, according to the number of times that the drop-down prompting result under each input prefix is selected, statistics obtains the weight of the drop-down prompting result under each input prefix, according to two kinds of weights that above-mentioned statistics obtains, calculate the probability that the drop-down prompting result under each input prefix is recommended, then by drop-down prompting result according to the probability arranged in sequence recommended.Due to the probability obtaining recommending according to the weight calculation of the drop-down prompting result under the weight of each input prefix and each input prefix, consider search custom and the search experience of user, make drop-down prompting result arranged in sequence more can meet the search intention of user, solve the drop-down prompting inputting prefix based on user in background technology thus and meet the low problem of the accuracy of user search intent, achieve the beneficial effect improving the drop-down prompting accuracy inputting prefix based on user.And, according to weight and the occurrence number of himself of all son input prefixes of input prefix, calculate the weight of input prefix, consider in all input prefix branches at input prefix place, the weight of its son input prefix, improves the accuracy of the weight of input prefix.

Further, in the process of the probability obtaining recommending according to the weight calculation counted on, go out to input the transition probability between prefix by the weight calculation counted on, and input prefix is to the transition probability of the drop-down prompting result under input prefix; Again according to weight and the corresponding transition probability of drop-down prompting result, calculate the probability that drop-down prompting result is selected under corresponding input prefix; Then each drop-down prompting result selected probability under different input prefixes is added up mutually, obtain the probability that each drop-down prompting result is recommended.The probability of transition probability, selected probability and recommendation is calculated successively according to weight, according to the weight that objective statistical obtains, and the impact input input of prefix and drop-down prompting result are by the objective law selected, COMPREHENSIVE CALCULATING obtains the probability recommended, and improves the drop-down prompting accuracy inputting prefix based on user on the whole.Above-mentioned explanation is only the general introduction of technical solution of the present invention, in order to technological means of the present invention can be better understood, and can be implemented according to the content of instructions, and can become apparent, below especially exemplified by the specific embodiment of the present invention to allow above and other objects of the present invention, feature and advantage.

Accompanying drawing explanation

By reading hereafter detailed description of the preferred embodiment, various other advantage and benefit will become cheer and bright for those of ordinary skill in the art.Accompanying drawing only for illustrating the object of preferred implementation, and does not think limitation of the present invention.And in whole accompanying drawing, represent identical parts by identical reference symbol.In the accompanying drawings:

Fig. 1 shows the drop-down prompting schematic diagram based on input prefix in prior art;

Fig. 2 shows search procedure inforamtion tree schematic diagram of the present invention;

Fig. 3 shows a kind of according to an embodiment of the invention drop-down reminding method process flow diagram based on input prefix;

Fig. 4 shows a kind of according to an embodiment of the invention drop-down reminding method process flow diagram based on input prefix;

Fig. 5 shows a kind of according to an embodiment of the invention based on drop-down hints model schematic diagram in the drop-down reminding method of input prefix;

Fig. 6 shows a kind of according to an embodiment of the invention schematic diagram based on user's input prefix in the drop-down reminding method of input prefix;

Fig. 7 shows the schematic diagram selecting drop-down prompting result in a kind of according to an embodiment of the invention drop-down reminding method based on input prefix;

Fig. 8 shows a kind of according to an embodiment of the invention drop-down suggestion device structural drawing based on input prefix;

Fig. 9 shows a kind of according to an embodiment of the invention drop-down suggestion device structural drawing based on input prefix;

Figure 10 shows a kind of according to an embodiment of the invention drop-down prompt system schematic diagram based on input prefix.

Embodiment

Below with reference to accompanying drawings exemplary embodiment of the present disclosure is described in more detail.Although show exemplary embodiment of the present disclosure in accompanying drawing, however should be appreciated that can realize the disclosure in a variety of manners and not should limit by the embodiment set forth here.On the contrary, provide these embodiments to be in order to more thoroughly the disclosure can be understood, and complete for the scope of the present disclosure can be conveyed to those skilled in the art.

The present invention can be applied to computer system/server, and it can operate with other universal or special computing system environment numerous or together with configuring.The example of the well-known computing system being suitable for using together with computer system/server, environment and/or configuration includes but not limited to: personal computer system, server computer system, thin client, thick client computer, hand-held or laptop devices, system based on microprocessor, Set Top Box, programmable consumer electronics, NetPC Network PC, minicomputer system large computer system and comprise the distributed cloud computing technology environment of above-mentioned any system, etc.

Computer system/server can describe under the general linguistic context of the computer system executable instruction (such as program module) performed by computer system.Usually, program module can comprise routine, program, target program, assembly, logic, data structure etc., and they perform specific task or realize specific abstract data type.Computer system/server can be implemented in distributed cloud computing environment, and in distributed cloud computing environment, task is performed by the remote processing devices by communication network links.In distributed cloud computing environment, program module can be positioned at and comprise on the Local or Remote computing system storage medium of memory device.

When the keyword of inputted search, user is in search column or other positions with function of search, after inputting a character, namely after inputting an input prefix, two kinds of situations can be there are, a kind of situation be from search column or have function of search other positions drop-down prompting select a drop-down prompting result consistent with the search intention of user directly to search for; Another situation continues input next input prefix, then checks in the list of drop-down prompting the drop-down prompting result whether listing needs.

An inputted search process of user can represent with one tree.The root node wherein set is first input prefix that user inputs, as shown in Figure 2.First input prefix that " feelings " in Fig. 2 input for user is the root node of tree; After user inputs first input prefix, then next input prefix " companion ", " interest ", " people " etc. may be inputted, input prefix " companion ", " interest ", " people " etc. constitute the second layer node of tree, the representative of this node layer be cumulative from the root node of tree to the path word this node layer or word, as " lovers " in Fig. 2, " temperament and interest ", " sweet heart " etc.

The root node of tree is first input prefix that user inputs; The leaf node of tree is the complete search word that user inputs; Also may there is intermediate node between root node and leaf node, intermediate node accumulated path word or word on root node obtain.

A kind of based on the input drop-down reminding method of prefix, device and system below by what enumerate that several specific embodiment introduces that the present invention proposes in detail.

Embodiment one, introduces a kind of drop-down reminding method based on input prefix that the present invention proposes in detail.

With reference to Fig. 3, show a kind of according to an embodiment of the invention drop-down reminding method process flow diagram based on input prefix.

Step S11, determines current input prefix.

, can there is drop-down prompting result according to the character of input in input character in the search column of the search column of user in webpage or browser, drop-down prompting result is all interrelated with the character inputted.Generally, the character of input is first or former characters of drop-down prompting result, or is first or former characters of first word or word in drop-down prompting result.

When the character inputted be need the character of search or need character a part of of search time, the character of described input can be understood as input prefix.

Such as, when user time input character " h ", determines that " h " is for current input prefix in search column.Character " h " can be the input prefix of the character such as " htc " or " hello ".If user continues to have input character " t ", so current input prefix then becomes " ht ".

Step S12, adds up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix.

After determining current input prefix in step s 11, with current input prefix for root, several even dozens of input prefix branch can be set up.For the current input prefix of root is expressed as root input prefix in input prefix branch.Other input prefix can also be comprised in input prefix branch, but other input prefix all with current input prefix for root inputs prefix; Last input prefix in each input prefix branch is leaf input prefix.

Wherein, each input prefix in described input prefix branch comprises at least successively:

Root input prefix and leaf input prefix; Described input prefix branch at least inputs prefix and one and this root by root and inputs leaf that prefix is associated and input prefix and form.

Such as, " h " is current input prefix, and the input prefix branch being root with " h " has two, and Article 1 is " h-ht-htc ", and Article 2 is " h-he ".Wherein, " h ", " ht ", " htc " are each input prefix in Article 1 input prefix branch, and " h " is root input prefix, and " htc " is leaf input prefix; " h " and " he " is each input prefix in Article 2 input prefix branch, and " h " is root input prefix, and " he " is leaf input prefix.

When after each input prefix of input, under the input prefix of input, there will be drop-down prompting result.

Such as, when by search column input " h " after, there will be " Korea S buys on behalf ", " with " etc. drop-down prompting result, then " Korea S buys on behalf ", " with " for root input prefix " h " drop-down prompting result; After input " htc ", there will be the drop-down prompting result such as " htc g12 " and " htc g11 " in search column, then " htc g12 " and " htc g11 " is the drop-down prompting result of leaf input prefix " htc ".

So in two the input prefix branches being root with " h ", " h ", " ht ", " htc " and " he " are each input prefix; " Korea S buys on behalf ", " with ", " htc g12 " and " htc g11 " be input prefix under drop-down prompting result.

In the above example, the weight of statistics input prefix " h ", " ht ", " htc " and " he "; Add up drop-down prompting result " Korea S buys on behalf ", " with ", the weight of " htc g12 " and " htc g11 ".

Step S13, according to the weight of described each input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended.

According to the weight of each input prefix counted in step S12, as inputted the weight of prefix " h ", " ht ", " htc " and " he ", with the weight of the drop-down prompting result under each input prefix, as drop-down prompting result " Korea S buys on behalf ", " with ", the weight of " htc g12 " and " htc g11 ", COMPREHENSIVE CALCULATING obtains the probability that the drop-down prompting result under each input prefix is recommended, as COMPREHENSIVE CALCULATING obtain drop-down prompting result " Korea S buys on behalf ", " with ", " htc g12 " and " htc g11 " probability of recommending.

Step S14, by the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented on the search column of client as the candidate item of current input prefix.

By probability arranged in sequence in drop-down prompting the results list that drop-down prompting result is recommended according to the drop-down prompting result calculated in step S13, generally, can according to the probability descending sort recommended, as drop-down prompting result " Korea S buys on behalf ", " with ", the probability that " htc g12 " and " htc g11 " recommends is respectively 0.04, 0.03, 0.02 and 0.01, then after input input prefix " h ", can according to " Korea S buys on behalf " in drop-down prompting the results list, " with ", the order of " htc g12 " and " htc g11 ", from top to bottom be arranged in order.

Wherein, the probability that drop-down prompting result " Korea S buys on behalf " is recommended is 0.04, maximum in the probability that each drop-down prompting result is recommended, then can using the candidate item of drop-down prompting result " Korea S buys on behalf " as current input prefix " h ", display in the client, as in the search column that is presented at browser or address field.

In sum, a kind of drop-down reminding method based on input prefix that the embodiment of the present invention one proposes, compared with prior art, has the following advantages:

A kind of drop-down reminding method based on input prefix that the embodiment of the present invention one proposes, in the input prefix branch being root with the current input prefix determined, statistics obtains the weight of each input prefix, with the weight of the drop-down prompting result under each input prefix, according to two kinds of weights that above-mentioned statistics obtains, calculate the probability that the drop-down prompting result under each input prefix is recommended, then by drop-down prompting result according to the probability arranged in sequence recommended.Due to the probability obtaining recommending according to the weight calculation of the drop-down prompting result under the weight of each input prefix and each input prefix, consider search custom and the search experience of user, make drop-down prompting result arranged in sequence more can meet the search intention of user, solve the drop-down prompting inputting prefix based on user in background technology thus and meet the low problem of the accuracy of user search intent, achieve the beneficial effect improving the drop-down prompting accuracy inputting prefix based on user.

Embodiment two, introduces a kind of drop-down reminding method based on input prefix that the present invention proposes in detail.

With reference to Fig. 4, show a kind of according to an embodiment of the invention drop-down reminding method process flow diagram based on input prefix.

Step S21, determines current input prefix.

Input character in search column in the softwares such as the search column of user in webpage or browser, when the character inputted be need the character of search or need the character of search a part of time, the character of described input can be understood as input prefix.

Search column or there is function of search the position such as address field in input character time, alphabet that is actual, that input is current input prefix.

Described character can comprise letter and word etc., does not limit the concrete scope of character herein, as long as the character belonged to for searching for can be defined as inputting prefix.

Such as, when user search column or there is function of search address field in input character " i " time, determine that " i " is for current input prefix.Character " i " can be the input prefix of the character such as " ipad " or " iphone "; As user's input " ip ", " ip " is current input prefix

Step S22, adds up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix.

After determining current input prefix in the step s 21, with current input prefix for root, several even dozens of input prefix branch can be set up.Each input prefix branch for root, presents tree-like or reticulate texture with current input prefix, for the current input prefix of root is expressed as root input prefix in input prefix branch; Last input prefix in each input prefix branch is leaf input prefix; Root input prefix in a certain input prefix branch can for obtaining leaf input prefix in other input prefix branches, as, in input prefix branch " hot-hotel ", " hot " is root input prefix, in input prefix branch " h-ho-hot ", " hot " is leaf input prefix.

Each input prefix in described input prefix branch can also comprise: intermediate input prefix;

Described intermediate input prefix is the input prefix between root input prefix and leaf input prefix.

Described intermediate input prefix can be multiple input prefix.

Intermediate input prefix in a certain input prefix branch can be also the root input prefix in other input prefix branches or leaf input prefix, and as in input prefix branch " h-hot-hotel ", " hot " is intermediate input prefix; In input prefix branch " h-ho-hot ", " hot " is leaf input prefix; In input prefix " hot-hotel ", " hot " is root input prefix.

Such as, " h " is current input prefix, and the input prefix branch being root with " h " has two, and Article 1 is " h-ht-htc ", and Article 2 is " h-he ".Wherein, " h ", " ht ", " htc " are each input prefix in Article 1 input prefix branch, and " h " is root input prefix, and " htc " is leaf input prefix, " ht ", between root input prefix " h " and leaf input prefix " htc ", " ht " is intermediate input prefix; " h " and " he " is each input prefix in Article 2 input prefix branch, and " h " is root input prefix, and " he " is leaf input prefix.

Such as, input " h " after, there will be " Korea S buys on behalf ", " with " etc. drop-down prompting result, then " Korea S buys on behalf ", " with " for root input prefix " h " drop-down prompting result; After input " ht ", there will be drop-down prompting results such as " htc hd ", then " htc hd " the drop-down prompting result that is intermediate input prefix " ht "; After input " htc ", there will be the drop-down prompting result such as " htc g12 " and " htc g11 ", then " htc g12 " and " htc g11 " is the drop-down prompting result of leaf input prefix " htc ".

In the above example, the weight of statistics input prefix " h ", " ht ", " htc " and " he "; Add up drop-down prompting result " Korea S buys on behalf ", " with ", " htc g12 ", " htc hd " and " htc g11 " weight.

In step S22, the weight (step S221 and step S222) of statistics input prefix belongs to coordination with the weight (step S223) of adding up the drop-down prompting result inputted under prefix, and step S22 specifically can comprise:

Step S221, the occurrence number of statistics root input prefix, leaf input prefix and intermediate input prefix.

According to the occurrence number of input prefix each in log statistic input prefix branch, namely statistics inputs the occurrence number of root input prefix, leaf input prefix and intermediate input prefix in prefix branch.

Described occurrence number is the number of times of user's input prefix.As user inputs " h " 50 times, then the occurrence number inputting prefix " h " is 50.

Such as, add up above-mentioned with in " h " two the input prefix branches that are root, the occurrence number of root input prefix " h " is 200, and the occurrence number of intermediate input prefix " ht " is 100, and the occurrence number of leaf input prefix " htc " and " he " is respectively 300 and 50.

Step S222, inputs the weight of occurrence number as this leaf input prefix of prefix using leaf; By each son input weight of prefix of intermediate input prefix and the occurrence number sum of intermediate input prefix, as the weight of this intermediate input prefix; Root is inputted the weight of each son input prefix and the occurrence number sum of root input prefix of prefix, as the weight of root input prefix.

Wherein, each son input prefix of described input prefix refers to that inputting the child that prefix has a set membership with root inputs prefix; Each son input prefix of described intermediate input prefix refers to that the child with intermediate input prefix with set membership inputs prefix.

Such as, the son input prefix of root input prefix " h " is intermediate input prefix " ht " and leaf input prefix " he "; The son input prefix of intermediate input prefix " ht " is leaf input prefix " htc ".

After the occurrence number of each input prefix counted in step S221, leaf is inputted the weight of occurrence number 300 as leaf input prefix " htc " of prefix " htc ", namely the weight of leaf input prefix " htc " is 300; In like manner, the weight of leaf input prefix " he " is 50; By the weight 300 of the son of intermediate input prefix " ht " input prefix " htc " and the occurrence number 100 sum 300+100=400 of intermediate input prefix " ht ", as the weight of intermediate input prefix " ht "; Root is inputted the occurrence number sum 50+400+200=650 of son input prefix " he " of prefix " h ", the weight of " ht " and root input prefix " h ", as the weight of root input prefix " h ".

That is, the weight of root input prefix " h " is 650, and the weight of intermediate input prefix " ht " is 400, and the weight of leaf input prefix " htc " is 300, and the weight of leaf input prefix " he " is 50.

Step S223, adds up the number of times that the drop-down prompting result under each input prefix is hit, and the number of times drop-down prompting result under each input prefix be hit is as the weight of the drop-down prompting result under each input prefix.

According to the number of times that the drop-down prompting result under input prefix each in log statistic input prefix branch is hit, the number of times be hit with the drop-down prompting result under intermediate input prefix under namely in statistics input prefix branch under root input prefix, leaf input prefix.

The number of times that described drop-down prompting result is hit, for after user's input prefix, selects the number of times of drop-down prompting result in drop-down prompting list.After user's input " h ", the number of times selecting drop-down prompting result " Korea S buys on behalf " in drop-down prompting list is 20 times, then the number of times that be hit of drop-down prompting result " Korea S buys on behalf " under input prefix " h " is 20.

Such as, add up above-mentioned with in " h " two the input prefix branches that are root, the number of times that drop-down prompting result " Korea S buys on behalf " is hit under input prefix " h " is 20 times; Drop-down prompting result " with " number of times that is hit under input prefix " h " is 25 times; The number of times that drop-down prompting result " htc hd " is hit under input prefix " ht " is 100 times; The number of times that drop-down prompting result " htc g12 " is hit under input prefix " htc " is 500 times.

Then the weight of drop-down prompting result " Korea S buys on behalf " under input prefix " h " is 20; Drop-down prompting result " with " weight under input prefix " h " is 25; The weight of drop-down prompting result " htc hd " under input prefix " ht " is 100; The weight of drop-down prompting result " htc g12 " under input prefix " htc " is 500.

Step S23, according to the weight of described each input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended.

In step S23, calculating transition probability (step S231) between input prefix with calculating inputs prefix and belongs to coordination to the transition probability (step S232) of the drop-down prompting result inputted under prefix, and step S23 specifically can comprise:

Step S231, inputs the transition probability between prefix according to the weight calculation of described each input prefix.

Transition probability between described input prefix, for after a certain input prefix of input, follows the probability of next input prefix of input.Such as, after input input prefix " h ", the probability of input input prefix " ht " is the transition probability between input prefix " h " to input prefix " ht "; After input input prefix " h ", the probability of input input prefix " htc " is the transition probability between input prefix " h " to input prefix " htc ".

Transition probability between described input prefix can be in same input prefix branch, root input prefix is to the transition probability of intermediate input prefix, intermediate input prefix, also can for root input prefix be to the transition probability of leaf input prefix to the transition probability of leaf input prefix.

To calculate in above-mentioned example, root input prefix " h " is that example is described to the transition probability of intermediate input prefix " ht ".

Can carry out according to following two steps according to the transition probability that the weight calculation of described each input prefix inputs between prefix:

A1) difference of the input weight of prefix and the occurrence number of this input prefix, is calculated;

The difference that the weight 650 of calculating root input prefix " h " and root input the occurrence number 200 of prefix " h " is 650-200=450;

A2), by a certain son of described input prefix input the weight of prefix and the described difference that should input prefix is divided by, obtaining the transition probability of this input prefix to this son input prefix;

The weight 400 root being inputted son input prefix intermediate input prefix " ht " of prefix " h " is divided by with the described difference 450 that root inputs prefix " h ", obtains the transition probability 400/450=0.889 of root input prefix " h " to intermediate input prefix " ht ".

Described input prefix comprises root input prefix and intermediate input prefix;

When there is one or several intermediate input prefix between root input prefix and a certain input prefix, root is inputted between prefix and described a certain input prefix, transition probability between input prefix to son input prefix is multiplied, and obtains the transition probability of root input prefix to described a certain input prefix.

Such as, calculate the transition probability of root input prefix " h " to leaf input prefix " htc ", need first to calculate the transition probability M1 of root input prefix " h " to intermediate input prefix " ht ", with the transition probability M2 of intermediate input prefix " ht " to leaf input prefix " htc ", again M1 and M2 is multiplied, obtains the transition probability M1xM2 of root input prefix " h " to leaf input prefix " htc ".

Step S232, according to the weight of the drop-down prompting result under described each input prefix and the weight of described input prefix, calculates the transition probability of input prefix to the drop-down prompting result under input prefix.

Described input prefix to input prefix under drop-down prompting result transition probability for input a certain input prefix after, the probability that the drop-down prompting result under this input prefix is hit.Such as, after input input prefix " h ", the probability that the drop-down prompting result " Korea S buys on behalf " under this input prefix " h " is hit is the transition probability of input prefix " h " to the drop-down prompting result " Korea S buys on behalf " under this input prefix " h ".

The weight of drop-down prompting result under input prefix and the weight of this input prefix are divided by, obtain the transition probability of this input prefix to the drop-down prompting result of this under this input prefix.

Such as, the transition probability of root input prefix " h " to the drop-down prompting result " Korea S buys on behalf " under root input prefix " h " is calculated.

The weight 20 of drop-down prompting result " Korea S buys on behalf " under root input prefix " h " is divided by with the weight 650 that root inputs prefix " h ", obtains the transition probability 20/650=0.031 of root input prefix " h " to the drop-down prompting result " Korea S buys on behalf " under root input prefix " h ".

Step S233, according to the weight of input prefix and the weight of the drop-down prompting result under input prefix, or according to the transition probability between described input prefix and the input prefix transition probability to the drop-down prompting result inputted under prefix, calculate the probability that drop-down prompting result under each input prefix is selected.

According to concrete input prefix, comprise root input prefix, intermediate input prefix and leaf input prefix, by probability selected for the drop-down prompting result under each input prefix of calculating, be divided into following three kinds of situations:

B1), when inputting prefix and being root input prefix, the weight that weight root being inputted the drop-down prompting result under prefix and root input prefix is divided by, and obtains root and inputs the selected probability of drop-down prompting result under prefix.

When calculating the selected probability of drop-down prompting result under root input prefix, such as, the probability that drop-down prompting result " Korea S buys on behalf " under calculating root input prefix " h " is selected, the weight 20 of drop-down prompting result " Korea S buys on behalf " under root input prefix " h " is divided by with the weight 650 that root input prefix " h ", obtains root and input the selected probability 20/650=0.031 of drop-down prompting result " Korea S buys on behalf " under prefix " h ".

B2), when input prefix is intermediate input prefix, root is inputted prefix to be multiplied with the transition probability of this intermediate input prefix to the drop-down prompting result under this intermediate input prefix to the transition probability of this intermediate input prefix, obtain the probability that this drop-down prompting result under this intermediate input prefix is selected.

When calculating the selected probability of drop-down prompting result under intermediate input prefix, such as, the probability that drop-down prompting result " htc hd " under calculating intermediate input prefix " ht " is selected, root is inputted prefix " h " to be multiplied to the transition probability M1=0.889 of intermediate input prefix " ht " with the transition probability M3 of intermediate input prefix " ht " to the drop-down prompting result " htc hd " under intermediate input prefix " ht ", the probability obtaining drop-down prompting result " htc hd " under intermediate input prefix " ht " selected is M1xM3.

B3), when inputting prefix and being leaf input prefix, root is inputted prefix to be multiplied with the transition probability that this leaf inputs the drop-down prompting result that prefix inputs under prefix to this leaf to the transition probability of this leaf input prefix, obtain this leaf and input the selected probability of this drop-down prompting result under prefix.

When calculating the selected probability of drop-down prompting result under leaf input prefix, such as, the probability that drop-down prompting result " htc g12 " under calculating leaf input prefix " htc " is selected, root is inputted prefix " h " to be multiplied with the transition probability M4 that leaf inputs the drop-down prompting result " htc g12 " that prefix " htc " inputs under prefix " htc " to leaf to the transition probability M1xM2 of leaf input prefix " htc ", obtain leaf and input the selected probability M1xM2xM4 of drop-down prompting result " htc g12 " under prefix " htc ".

Step S234, the probability recommended according to described selected probability calculation.

Probability selected under each input prefix for same drop-down prompting result is added, obtains the probability that this drop-down prompting result is recommended.

Such as, drop-down prompting result " htc g13 " selected probability under root input prefix " h " is N1, probability selected under intermediate input prefix " ht " is N2, under leaf input prefix " htc ", selected probability is N3, then N1, N2 and N3 are added, obtain the probability N1+N2+N3 that drop-down prompting result " htc g13 " is recommended.

Step S24, by the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented on the search column of client as the candidate item of current input prefix.

By probability arranged in sequence in drop-down prompting the results list that drop-down prompting result is recommended according to the drop-down prompting result calculated in step S234, generally, can according to the probability descending sort recommended, as drop-down prompting result " Korea S buys on behalf ", " with ", the probability that " htc g12 " and " htc g11 " recommends is respectively 0.04, 0.03, 0.02 and 0.01, then after input input prefix " h ", can according to " Korea S buys on behalf " in drop-down prompting the results list, " with ", the order of " htc g12 " and " htc g11 ", from top to bottom be arranged in order.

Embodiment three, introduces a kind of drop-down reminding method based on input prefix that the present invention proposes in detail.

With reference to Fig. 5, show a kind of according to an embodiment of the invention based on drop-down hints model schematic diagram in the drop-down reminding method of input prefix.

Fig. 5 describes with " h " the drop-down hints model for root input prefix.Wherein, " h " is root input prefix, and " ht " is intermediate input prefix, and " he " and " htc " is leaf input prefix." Korea S buys on behalf ", " htc g13 ", " wedding gauze kerchief ", " with ", " htc hd2 ", " htc g12 ", " htc g7 " and " htc " (this is drop-down prompting result, be different from leaf input prefix) be drop-down prompting result.And indicate occurrence number and the weight of each input prefix, and the number of times that drop-down prompting result is hit under difference input prefix.

With reference to Fig. 6, show a kind of according to an embodiment of the invention schematic diagram based on user's input prefix in the drop-down reminding method of input prefix.

First input " h ", then " h " is as root input prefix, after input " h ", also may input " t " or " e ", if that input " h " input afterwards is " t ", then next also may input " c ".

With reference to Fig. 7, show the schematic diagram selecting drop-down prompting result in a kind of according to an embodiment of the invention drop-down reminding method based on input prefix.

User input " h " after, " Korea S buys on behalf " may be selected from drop-down prompting result, also may select " htc g13 ", " wedding gauze kerchief " or " with "; " htc g13 ", " htc " or " htc hd2 " after inputting " t " again, may be selected; " htc g13 ", " htc g12 " or " htc g7 " after inputting " c " again, may be selected.

The occurrence number of statistics root input prefix " h " is 100, the occurrence number of intermediate input prefix " he " is 223, the occurrence number of intermediate input prefix " ht " is 145, the occurrence number of leaf input prefix " htc " is 2156, then the weight of leaf input prefix " htc " is 2156, the weight of intermediate input prefix " ht " is 2156+145=2301, and the weight of root input prefix " h " is 100+223+145+2156=2624.

Calculating root input prefix " h " is p (ht|h)=2301/ (2624-100)=0.9 to the transition probability of intermediate input prefix " ht "; Calculating intermediate input prefix " ht " is p (htc|ht)=2156/ (2301-145)=1 to the transition probability of leaf input prefix " htc "; Root input prefix " h " is P (htc|h)=p (ht|h) * p (htc|ht)=0.9*1=0.9 to the transition probability of leaf input prefix " htc ".

The input prefix of the current input of user is " h ", then the selected probability of the drop-down prompting result under root input prefix " h " is:

P (Korea S buys on behalf | and h)=0.03

p(htcg13|h)=0.02

P (wedding gauze kerchief | h)=0.02

P (with | h)=0.003

The input prefix of the current input of user is " h ", may continue afterwards to input " t ", then the selected probability of the drop-down prompting result under intermediate input prefix " ht " is:

p(ht|h)*p(htcg13|ht)=0.9*0.03=0.027

p(ht|h)*p(htc|ht)=0.9*0.02=0.018

p(ht|h)*p(|htc hd2|ht)=0.9*0.003=0.003

The input prefix of the current input of user is " h ", may continue afterwards to input " t " and " c ", then the probability that leaf inputs drop-down prompting result under prefix " htc " selected is:

p(ht|h)*p(htc|ht)*p(htcg13|htc)=0.9*1*0.45=0.405

p(ht|h)*p(htc|ht)*p(htc g12|htc)=0.9*1*0.3=0.27

p(ht|h)*p(htc|ht)*p(htc g7|htc)=0.9*1*0.23=0.207

The probability that each drop-down prompting result of final calculating is recommended:

P (Korea S buys on behalf)=p (Korea S buys on behalf | and h)=0.03

P (wedding gauze kerchief)=p (wedding gauze kerchief | h)=0.02

P (with)=p (with | h)=0.003

p(htc)=p(ht|h)*p(htc|ht)=1*0.02=0.02

p(htc hd2)=p(ht|h)*p(htc hd2|ht)==0.9*0.003=0.003

p(htc g12)=p(ht|h)*p(htc|ht)*p(htc g12|htc)=0.9*1*0.3=0.27

p(htc g7)=p(ht|h)*p(htc|ht)*p(htc g7|htc)=0.9*1*0.23=0.207

The probability size sequence recommended according to each drop-down prompting result, chooses the candidate of the large input prefix as active user's input of probability.

When after user's input prefix " h ", drop-down prompting result put in order for " htc g13 "-" Korea S buys on behalf "-" wedding gauze kerchief "-" with ".

When after user's input prefix " ht ", putting in order as " htc g13 "-" htc "-" htc hd2 " of drop-down prompting result.

When after user's input prefix " htc ", putting in order as " htc g13 "-" htc g7 "-" htc g12 " of drop-down prompting result.

Based on above-described embodiment, for Fig. 1, if use the method that the embodiment of the present invention provides, user inputs " h ", and input again " o " after " h ", if in conjunction with input and the search custom of user, before " hotmail " may come " hold lives to like ", but the result " hold lives to like " according to PV sequence in background technology comes before " hotmail ".

It can thus be appreciated that, a kind of drop-down reminding method based on input prefix proposed in the embodiment of the present invention three, objective, integrally to consider user search custom and search experience, make drop-down prompting result arranged in sequence more can meet the search intention of user, improve the drop-down prompting accuracy inputting prefix based on user.

In the present embodiment three, user inputs " h ", determines " h " for current input prefix, with " h " for root constructs two inputs prefix branch " h-e " and " h-t-c ".Wherein, the son input prefix that " he " and " ht " is root input prefix " h ", according to the weight of all sons input prefix " he " and " ht " of root input prefix " h ", the weight of root input prefix " h " is calculated with the occurrence number of root input prefix " h ", take into account in all input prefix branches at root input prefix " h " place, the weight of its all son input prefix, improves the accuracy of the weight of input prefix.

And, according to root input prefix " h " that objective statistical obtains, intermediate input prefix " ht ", the weight of drop-down prompting result under the weight of leaf input prefix " htc " and each input prefix, and the input of impact input prefix (as, may input " t " after input " h ", if input " t " also may input " c " afterwards, represented by the transition probability between input prefix, comprise the transition probability of root input prefix " h " to intermediate input prefix " ht ", intermediate input prefix " ht " is to the transition probability of leaf input prefix " htc ", with the transition probability of root input prefix " h " to leaf input prefix " htc ") and drop-down prompting result selected the objective law of (probability selected by drop-down prompting result represents), COMPREHENSIVE CALCULATING obtains the probability that each drop-down prompting result is recommended, improve the drop-down prompting accuracy inputting prefix based on user on the whole.

If that user's input is not " h ", but " ht ", then " ht " is current input prefix, builds input prefix branch, can obtain input prefix branch " ht-htc " with " ht " for root; In like manner, as user's input " htc ", then " htc " is current input prefix, builds input prefix branch, can obtain input prefix branch " htc-htcds-htcds2 " with " htc " for root.The step of the probability that subsequent calculations drop-down prompting result is recommended is also identical with the principle introduced in above-described embodiment, no longer repeats at this.

If user's input alphabet " l ", then " l " is current input prefix, with " l " for the input prefix that root builds branches into " l-o-o-k " and " l-u-c-k ".Wherein, the probability recommended of the drop-down prompting result under each input prefix in advance off-line analysis calculate.As user's input " l ", directly root is inputted the probability sorting display of the recommendation that the drop-down prompting result under prefix " l " calculates according to respective off-line analysis in advance; As user's input " lu ", directly root is inputted the probability sorting display of the recommendation that the drop-down prompting result under prefix " lu " calculates according to respective off-line analysis in advance.

A kind of drop-down reminding method based on input prefix that the present embodiment three proposes, a large amount of input prefix branches is constructed according to daily record, and off-line analysis calculates the probability that each drop-down prompting result in each input prefix branch under each input prefix is recommended in advance, when user's input input prefix for described in input prefix in a large amount of input prefix branches of constructing time, directly the probability sorting that the drop-down prompting result in advance under this input prefix of calculating of off-line analysis is recommended is presented on the search column of client.

Embodiment four, introduces a kind of drop-down suggestion device based on input prefix that the present invention proposes in detail.

With reference to Fig. 8, show a kind of according to an embodiment of the invention drop-down suggestion device structural drawing based on input prefix.

Described device comprises:

Current input prefix determination module 41, weight statistical module 42, recommends probability evaluation entity 43, and, sequencing display module 44.

Introduce the relation between the function of each module and each module below respectively in detail.

Current input prefix determination module 41, is suitable for determining current input prefix.

The input prefix that described current input prefix determination module 41 inputs based on user, determines current input prefix.

Weight statistical module 42, is suitable for adding up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix.

In the input prefix branch that the current input prefix determined with described current input prefix determination module 41 is root, described weight statistical module 42 adds up the weight of each input prefix, and respectively inputs the weight of the drop-down prompting result under prefix.

Recommend probability evaluation entity 43, be suitable for the weight according to described each input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended.

The weight of each input prefix that described recommendation probability evaluation entity 43 is added up according to described weight statistical module 42, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended.

Sequencing display module 44, is suitable for the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented on the search column of client as the candidate item of current input prefix.

The probability arranged in sequence of the recommendation that drop-down prompting result calculates according to described recommendation probability evaluation entity 43 by described sequencing display module 44, and select to recommend the large drop-down prompting result of probability to be presented on the search column of client as the candidate item of current input prefix.

In sum, a kind of drop-down suggestion device based on input prefix that the embodiment of the present invention four proposes, compared with prior art, has the following advantages:

A kind of drop-down suggestion device based on input prefix that the embodiment of the present invention four proposes, in the input prefix branch being root with the current input prefix determined, statistics obtains the weight of each input prefix, with the weight of the drop-down prompting result under each input prefix, according to two kinds of weights that above-mentioned statistics obtains, calculate the probability that the drop-down prompting result under each input prefix is recommended, then by drop-down prompting result according to the probability arranged in sequence recommended.Due to the probability obtaining recommending according to the weight calculation of the drop-down prompting result under the weight of each input prefix and each input prefix, consider search custom and the search experience of user, make drop-down prompting result arranged in sequence more can meet the search intention of user, solve the drop-down prompting inputting prefix based on user in background technology thus and meet the low problem of the accuracy of user search intent, achieve the beneficial effect improving the drop-down prompting accuracy inputting prefix based on user.

Embodiment five, introduces a kind of drop-down suggestion device based on input prefix that the present invention proposes in detail.

With reference to Fig. 9, show a kind of according to an embodiment of the invention drop-down suggestion device structural drawing based on input prefix.

Described device comprises:

Current input prefix determination module 51, weight statistical module 52, recommends probability evaluation entity 53, and, sequencing display module 54.

Wherein, described weight statistical module 52 comprises:

Occurrence number statistics submodule 521, first weight determination submodule 522, second weight determination submodule the 523, three weight determination submodule 524, and, the 4th weight determination submodule 525.

Described recommendation probability evaluation entity 53 comprises:

First transition probability calculating sub module 531, second transition probability calculating sub module 532, chooses probability calculation submodule 533, and, recommend probability calculation submodule 534.

Introduce in detail respectively below each module and each submodule function and between relation.

Current input prefix determination module 51, is suitable for determining current input prefix.

Such as, when user is in search column time input character " h ", described current input prefix determination module 51 determines that " h " is for current input prefix.Character " h " can be the input prefix of the character such as " htc " or " hello ".

Weight statistical module 52, is suitable for adding up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix.

In the input branch that the current input prefix determined with described current input prefix determination module 51 is root, for the current input prefix of root is expressed as root input prefix in input prefix branch; Last input prefix in each input prefix branch is leaf input prefix.

In the above example, described weight statistical module 52 adds up the weight of input prefix " h ", " ht ", " htc " and " he "; Described weight statistical module 52 add up drop-down prompting result " Korea S buys on behalf ", " with ", " htc g12 ", " htc hd " and " htc g11 " weight.

Described weight statistical module 52, comprising:

Occurrence number statistics submodule 521, is suitable for the occurrence number of adding up root input prefix, leaf input prefix and intermediate input prefix.

Described occurrence number statistics submodule 521 adds up the occurrence number inputting each input prefix in prefix branch, and namely in statistics input prefix branch, root inputs the occurrence number of prefix, leaf input prefix and intermediate input prefix.

Such as, described occurrence number statistics submodule 521 adds up above-mentioned with in " h " two the input prefix branches that are root, the occurrence number of root input prefix " h " is 200, the occurrence number of intermediate input prefix " ht " is 100, and the occurrence number of leaf input prefix " htc " and " he " is respectively 300 and 50.

Described occurrence number statistics submodule 521, is also suitable for adding up the number of times that the drop-down prompting result under each input prefix is hit.

Described occurrence number statistics submodule 521 adds up the number of times that the drop-down prompting result in input prefix branch under each input prefix is hit, the number of times be hit with the drop-down prompting result under intermediate input prefix under namely in statistics input prefix branch under root input prefix, leaf input prefix.

Such as, described occurrence number statistics submodule 521 adds up above-mentioned with in " h " two the input prefix branches that are root, and the number of times that drop-down prompting result " Korea S buys on behalf " is hit under input prefix " h " is 20 times; Drop-down prompting result " with " number of times that is hit under input prefix " h " is 25 times; The number of times that drop-down prompting result " htc hd " is hit under input prefix " ht " is 100 times; The number of times that drop-down prompting result " htc g12 " is hit under input prefix " htc " is 500 times.

First weight determination submodule 522, the occurrence number being suitable for leaf to input prefix inputs the weight of prefix as this leaf.

Such as, leaf is inputted the weight of occurrence number 300 as leaf input prefix " htc " of prefix " htc " by described first weight determination submodule 522.

Second weight determination submodule 523, is suitable for each son input weight of prefix of intermediate input prefix and the occurrence number sum of intermediate input prefix, as the weight of this intermediate input prefix.

Such as, the weight 300 of described second weight determination submodule 523 by the son of intermediate input prefix " ht " input prefix " htc " and the occurrence number 100 sum 300+100=400 of intermediate input prefix " ht ", as the weight of intermediate input prefix " ht ".

Wherein, each son input prefix of described intermediate input prefix refers to that the child with intermediate input prefix with set membership inputs prefix.

Such as, the son input prefix of intermediate input prefix " ht " is leaf input prefix " htc ".

3rd weight determination submodule 524, is suitable for the weight of each son input prefix and the occurrence number sum of root input prefix that root are inputted prefix, as the weight of root input prefix.

Such as, root is inputted the occurrence number sum 50+400+200=650 of son input prefix " he " of prefix " h ", the weight of " ht " and root input prefix " h " by described 3rd weight determination submodule 524, as the weight of root input prefix " h ".

Wherein, each son input prefix of described input prefix refers to that inputting the child that prefix has a set membership with root inputs prefix.

Such as, the son input prefix of root input prefix " h " is intermediate input prefix " ht " and leaf input prefix " he ".

4th weight determination submodule 525, is suitable for the number of times that the drop-down prompting result under each input prefix the is hit weight as the drop-down prompting result under each input prefix.

Recommend probability evaluation entity 53, be suitable for the weight according to described each input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended.

Described recommendation probability evaluation entity 53, comprising:

First transition probability calculating sub module 531, the transition probability between being suitable for according to the weight calculation of described each input prefix input prefix.

Transition probability between described input prefix, for after a certain input prefix of input, follows the probability of next input prefix of input.

Described first transition probability calculating sub module 531 calculates the difference of the input weight of prefix and the occurrence number of this input prefix, the a certain son of described input prefix is inputted the weight of prefix and is divided by the described difference that should input prefix, obtains the transition probability of this input prefix to this son input prefix.

The difference that the weight 650 that described first transition probability calculating sub module 531 calculates root input prefix " h " and root input the occurrence number 200 of prefix " h " is 650-200=450.

The weight 400 that root is inputted son input prefix intermediate input prefix " ht " of prefix " h " by described first transition probability calculating sub module 531 is divided by with the described difference 450 that root inputs prefix " h ", obtains the transition probability 400/450=0.889 of root input prefix " h " to intermediate input prefix " ht ".

Described input prefix comprises root input prefix and intermediate input prefix.

When there is one or several intermediate input prefix between root input prefix and a certain input prefix, root inputs between prefix and described a certain input prefix by described first transition probability calculating sub module 531, transition probability between input prefix to son input prefix is multiplied, and obtains the transition probability of root input prefix to described a certain input prefix.

Such as, described first transition probability calculating sub module 531 calculates the transition probability of root input prefix " h " to leaf input prefix " htc ", described first transition probability calculating sub module 531 needs first to calculate the transition probability M1 of root input prefix " h " to intermediate input prefix " ht ", with the transition probability M2 of intermediate input prefix " ht " to leaf input prefix " htc ", again M1 and M2 is multiplied, obtains the transition probability M1xM2 of root input prefix " h " to leaf input prefix " htc ".

Second transition probability calculating sub module 532, is suitable for the weight of weight according to the drop-down prompting result under described each input prefix and described input prefix, calculates the transition probability of input prefix to the drop-down prompting result inputted under prefix.

Described input prefix to input prefix under drop-down prompting result transition probability for input a certain input prefix after, the probability that the drop-down prompting result under this input prefix is hit.

The weight of drop-down prompting result under input prefix and the weight of this input prefix are divided by by described second transition probability calculating sub module 532, obtain the transition probability of this input prefix to the drop-down prompting result of this under this input prefix.

Such as, the weight 20 of drop-down prompting result " Korea S buys on behalf " under root input prefix " h " is divided by with the weight 650 that root inputs prefix " h " by described second transition probability calculating sub module 532, obtains the transition probability 20/650=0.031 of root input prefix " h " to the drop-down prompting result " Korea S buys on behalf " under root input prefix " h ".

Choose probability calculation submodule 533, be suitable for according to inputting the weight of prefix and inputting the weight of the drop-down prompting result under prefix, or according to the transition probability between described input prefix and the input prefix transition probability to the drop-down prompting result inputted under prefix, calculate the probability that drop-down prompting result under each input prefix is selected.

When inputting prefix and being root input prefix, described in choose probability calculation submodule 533 root to be inputted the drop-down prompting result under prefix weight be divided by with the weight that root inputs prefix, obtain root and input the selected probability of drop-down prompting result under prefix.

Such as, the weight 20 of drop-down prompting result " Korea S buys on behalf " under root input prefix " h " is divided by with the weight 650 that root input prefix " h " by described probability calculation submodule 533 of choosing, and obtains root and inputs the selected probability 20/650=0.031 of drop-down prompting result " Korea S buys on behalf " under prefix " h ".

When input prefix is intermediate input prefix, describedly choose probability calculation submodule 533 root to be inputted prefix to be multiplied with the transition probability of this intermediate input prefix to the drop-down prompting result under this intermediate input prefix to the transition probability of this intermediate input prefix, obtain the probability that this drop-down prompting result under this intermediate input prefix is selected.

Such as, describedly choose probability calculation submodule 533 root to be inputted prefix " h " to be multiplied to the transition probability M1=0.889 of intermediate input prefix " ht " with the transition probability M3 of intermediate input prefix " ht " to the drop-down prompting result " htc hd " under intermediate input prefix " ht ", the probability obtaining drop-down prompting result " htc hd " under intermediate input prefix " ht " selected is M1xM3.

When inputting prefix and being leaf input prefix, describedly choose probability calculation submodule 533 root to be inputted prefix to be multiplied with the transition probability that this leaf inputs the drop-down prompting result that prefix inputs under prefix to this leaf to the transition probability of this leaf input prefix, obtain this leaf and input the selected probability of this drop-down prompting result under prefix.

Such as, describedly choose probability calculation submodule 533 root to be inputted prefix " h " to be multiplied with the transition probability M4 that leaf inputs the drop-down prompting result " htc g12 " that prefix " htc " inputs under prefix " htc " to leaf to the transition probability M1xM2 of leaf input prefix " htc ", obtain leaf and input the selected probability M1xM2xM4 of drop-down prompting result " htc g12 " under prefix " htc ".

Recommend probability calculation submodule 534, be suitable for the probability recommended according to described selected probability calculation.

Probability selected under each input prefix for same drop-down prompting result is added by described recommendation probability calculation submodule 534, obtains the probability that this drop-down prompting result is recommended.

Such as, drop-down prompting result " htc g13 " selected probability under root input prefix " h " is N1, probability selected under intermediate input prefix " ht " is N2, under leaf input prefix " htc ", selected probability is N3, then N1, N2 and N3 are added by described recommendation probability calculation submodule 534, obtain the probability N1+N2+N3 that drop-down prompting result " htc g13 " is recommended.

Sequencing display module 54, is suitable for the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented at client as the candidate item of current input prefix.

Probability arranged in sequence in drop-down prompting the results list that drop-down prompting result is recommended according to the drop-down prompting result that described recommendation probability calculation submodule 534 calculates by described sequencing display module 54.Generally, can according to the probability descending sort recommended.

And, the probability that drop-down prompting result " Korea S buys on behalf " is recommended is 0.04, maximum in the probability that each drop-down prompting result is recommended, described sequencing display module 54 can using the candidate item of drop-down prompting result " Korea S buys on behalf " as current input prefix " h ", display in the client, as in the search column that is presented at browser or address field.

A kind of drop-down suggestion device based on input prefix that the embodiment of the present invention five proposes, by module each in device and each submodule, in the input prefix branch being root with the current input prefix determined, according to the occurrence number of each input prefix, statistics obtains the weight of each input prefix, according to the number of times that the drop-down prompting result under each input prefix is selected, statistics obtains the weight of the drop-down prompting result under each input prefix, according to two kinds of weights that above-mentioned statistics obtains, calculate the probability that the drop-down prompting result under each input prefix is recommended, again by the probability arranged in sequence of drop-down prompting result according to recommendation.Due to the probability obtaining recommending according to the weight calculation of the drop-down prompting result under the weight of each input prefix and each input prefix, consider search custom and the search experience of user, make drop-down prompting result arranged in sequence more can meet the search intention of user, solve the drop-down prompting inputting prefix based on user in background technology thus and meet the low problem of the accuracy of user search intent, achieve the beneficial effect improving the drop-down prompting accuracy inputting prefix based on user.And, according to weight and the occurrence number of himself of all son input prefixes of input prefix, calculate the weight of input prefix, consider in all input prefix branches at input prefix place, the weight of its son input prefix, improves the accuracy of the weight of input prefix.

Further, in the process of the probability obtaining recommending according to the weight calculation counted on, go out to input the transition probability between prefix by the weight calculation counted on, and input prefix is to the transition probability of the drop-down prompting result under input prefix; Again according to weight and the corresponding transition probability of drop-down prompting result, calculate the probability that drop-down prompting result is selected under corresponding input prefix; Then each drop-down prompting result selected probability under different input prefixes is added up mutually, obtain the probability that each drop-down prompting result is recommended.The probability of transition probability, selected probability and recommendation is calculated successively according to weight, according to the weight that objective statistical obtains, and the impact input input of prefix and drop-down prompting result are by the objective law selected, COMPREHENSIVE CALCULATING obtains the probability recommended, and improves the drop-down prompting accuracy inputting prefix based on user on the whole.

Embodiment six, introduces a kind of drop-down prompt system based on input prefix that the present invention proposes in detail.

With reference to Figure 10, show a kind of according to an embodiment of the invention drop-down prompt system schematic diagram based on input prefix.

Described system in advance off-line analysis calculates the probability of each drop-down prompting result recommendation in each input prefix branch under each input prefix, when the input prefix that user inputs on the client for described in input prefix in a large amount of input prefix branches of constructing time, the probability directly will recommended according to the drop-down prompting result in advance under this input prefix of calculating of off-line analysis, the candidate item obtained that sorts is presented on the search column of client.

Described system comprises:

Client 61, agent node 62, meta data server 63, and, memory node 64.

Wherein, agent node 62, specifically can comprise:

Inquiry forwarding module 621, data forwarding module 622, routing table synchronization module 623, and, routing table update module 624.

Client 61, is suitable for receiving the request of search target keywords, and the request of described search target keywords is sent to agent node 62, comprise current input prefix in the request of described search target keywords.

Agent node 62, agent node 62 specifically can comprise:

Inquiry forwarding module 621, is suitable for described search target keywords request forward to corresponding memory node 64; Corresponding memory node is inquired about based on overall routing table information.

Data forwarding module 622, the candidate item of the current input prefix of the correspondence being suitable for memory node 64 to return is forwarded to client 61.

The candidate item of described correspondence current input prefix is the drop-down prompting result of recommending probability large.

Routing table synchronization module 623, is suitable for the overall routing table information of meta data server 63 to be synchronized to agent node 62.

Routing table update module 624, is suitable for, when receiving the routing table update notice of meta data server 63, upgrading overall routing table information from meta data server 63.

When there is certain memory node 64 fault or newly-increased memory node 64, the overall routing table information that meta data server 63 is safeguarded all can change, and thus, now can upgrade overall routing table information by routing table update module 624.

Meta data server 63, is suitable for safeguarding the overall routing table information pointing to memory node 64; Safeguard to store and revise accordingly and upgrade.

At least one memory node 64, be suitable for the client 61 inputted search target keywords request based on receiving, desired data is obtained from this locality stores, the probability that the drop-down prompting result of the current input prefix namely calculated according to off-line analysis is in advance recommended, sort the candidate item that draws, and send current input prefix candidate item to agent node 62.

In the system described in the present embodiment six, the inputted search target keywords request received is sent to agent node 62 by client 61.

Request forward, based on overall routing table information, is given corresponding memory node 64 by the inquiry forwarding module 621 in agent node 62.

Memory node 64, based on the request received, obtains desired data from this locality stores, and the response that transmission comprises target Value is wrapped to agent node 62.

The response bag of what memory node 64 returned by the data forwarding module 622 in agent node 62 comprise target Value sends client 61 to.

So far, described system completes once based on the course of work of the drop-down reminding method of input prefix.

It is emphasized that, the desired data that memory node 64 obtains from this locality stores, can be calculate the probability that each drop-down prompting result in each input prefix branch under each input prefix recommends, the candidate item of each input prefix drawn that sorts according to off-line analysis in advance.Memory node 64 can comprise above-mentioned a kind of based on all or part of module in the drop-down suggestion device of input prefix.That is, memory node 64 can comprise: current input prefix determination module, is suitable for determining current input prefix; Weight statistical module, is suitable for adding up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix; Recommend probability evaluation entity, be suitable for the weight according to described each input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended; Sequencing display module, is suitable for the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented at client as the candidate item of current input prefix; Wherein, each input prefix in described input prefix branch comprises at least successively: root input prefix and leaf input prefix; Described input prefix branch at least inputs prefix and one and this root by root and inputs leaf that prefix is associated and input prefix and form.

Just, the computation process obtaining in each input prefix branch the probability that each drop-down prompting result under each input prefix is recommended in advance off-line is carried out, when user inputs certain input prefix in client 61, by agent node 62, will input the probability of each drop-down prompting result recommendation in prefix branch under each input prefix according to this input prefix place that off-line analysis calculates in advance, the candidate item of this input prefix drawn that sorts is sent to display in client 61.Memory node 64 also can comprise the candidate item of corresponding current input prefix, the candidate item of described correspondence current input prefix is in the probability of each drop-down prompting result recommendation in each input prefix branch that off-line analysis calculates in advance under each input prefix, the drop-down prompting result that probability is large.

When routing table update notice is sent to agent node 62 by meta data server 63, the routing table update module 624 in agent node 62, obtains new overall routing table information from meta data server 63.

Further, routing table synchronization module 623, is synchronized to home agent node 62 by the overall routing table information of meta data server 63.

When client 61 receives search target keywords request in real time, the candidate item of the current input prefix of correspondence that off-line analysis in advance can be calculated loads and is shown on the search column of client 61.

If off-line analysis calculates two data in advance:

Input prefix K1 and its candidate item V1;

Input prefix K2 and its candidate item V2.

When client 61 real-time reception to the request of search target keywords in current input prefix be K1 time, what off-line analysis in advance can be calculated is that the candidate item V1 that K1 is corresponding is presented on the search column of client with current input prefix; When client 61 real-time reception to the request of search target keywords in current input prefix be K2 time, what off-line analysis in advance can be calculated is that the candidate item V2 that K2 is corresponding is presented on the search column of client with current input prefix.

Described system can upgrade one by one to the data that off-line analysis calculates in advance, and such as, the candidate item V1 corresponding to input prefix K1 carries out a renewal rewards theory; The candidate item V2 corresponding to input prefix K2 carries out a renewal rewards theory.Described renewal rewards theory can carry out by multithreading simultaneously, unaffected each other.

It should be noted that, for aforesaid embodiment of the method, in order to simple description, therefore it is all expressed as a series of combination of actions, but those skilled in the art should know, the present invention is not by the restriction of described sequence of movement, because according to the present invention, some step can adopt other orders or carry out simultaneously.Secondly, those skilled in the art also should know, the embodiment described in instructions all belongs to preferred embodiment, and involved action might not be essential to the invention.

For said apparatus embodiment, due to itself and embodiment of the method basic simlarity, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.

Each embodiment in this instructions all adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar part mutually see.

Those skilled in the art are easy to it is envisioned that: the combination in any application of each embodiment above-mentioned is all feasible, therefore the combination in any between each embodiment above-mentioned is all embodiment of the present invention, but this instructions does not just detail one by one at this as space is limited.

Intrinsic not relevant to any certain computer, virtual system or miscellaneous equipment with display at this algorithm provided.Various general-purpose system also can with use based on together with this teaching.According to description above, the structure constructed required by this type systematic is apparent.In addition, the present invention is not also for any certain programmed language.It should be understood that and various programming language can be utilized to realize content of the present invention described here, and the description done language-specific is above to disclose preferred forms of the present invention.

In instructions provided herein, describe a large amount of detail.But can understand, embodiments of the invention can be put into practice when not having these details.In some instances, be not shown specifically known method, structure and technology, so that not fuzzy understanding of this description.

Similarly, be to be understood that, in order to simplify the disclosure and to help to understand in each inventive aspect one or more, in the description above to exemplary embodiment of the present invention, each feature of the present invention is grouped together in single embodiment, figure or the description to it sometimes.But, the method for the disclosure should be construed to the following intention of reflection: namely the present invention for required protection requires feature more more than the feature clearly recorded in each claim.Or rather, as claims above reflect, all features of disclosed single embodiment before inventive aspect is to be less than.Therefore, the claims following embodiment are incorporated to this embodiment thus clearly, and wherein each claim itself is as independent embodiment of the present invention.

Those skilled in the art are appreciated that and adaptively can change the module in the equipment in embodiment and they are arranged in one or more equipment different from this embodiment.Module in embodiment or unit or assembly can be combined into a module or unit or assembly, and multiple submodule or subelement or sub-component can be put them in addition.Except at least some in such feature and/or process or unit be mutually repel except, any combination can be adopted to combine all processes of all features disclosed in this instructions (comprising adjoint claim, summary and accompanying drawing) and so disclosed any method or equipment or unit.Unless expressly stated otherwise, each feature disclosed in this instructions (comprising adjoint claim, summary and accompanying drawing) can by providing identical, alternative features that is equivalent or similar object replaces.

In addition, those skilled in the art can understand, although embodiments more described herein to comprise in other embodiment some included feature instead of further feature, the combination of the feature of different embodiment means and to be within scope of the present invention and to form different embodiments.Such as, in superincumbent claims, the one of any of embodiment required for protection can use with arbitrary array mode.

All parts embodiment of the present invention with hardware implementing, or can realize with the software module run on one or more processor, or realizes with their combination.It will be understood by those of skill in the art that the some or all functions based on the some or all parts in the input drop-down suggestion device of prefix and system that microprocessor or digital signal processor (DSP) can be used in practice to realize according to the embodiment of the present invention.The present invention can also be embodied as part or all equipment for performing method as described herein or device program (such as, computer program and computer program).Realizing program of the present invention and can store on a computer-readable medium like this, or the form of one or more signal can be had.Such signal can be downloaded from internet website and obtain, or provides on carrier signal, or provides with any other form.

The present invention will be described instead of limit the invention to it should be noted above-described embodiment, and those skilled in the art can design alternative embodiment when not departing from the scope of claims.In the claims, any reference symbol between bracket should be configured to limitations on claims.Word " comprises " not to be got rid of existence and does not arrange element in the claims or step.Word "a" or "an" before being positioned at element is not got rid of and be there is multiple such element.The present invention can by means of including the hardware of some different elements and realizing by means of the computing machine of suitably programming.In the unit claim listing some devices, several in these devices can be carry out imbody by same hardware branch.Word first, second and third-class use do not represent any order.Can be title by these word explanations.

Claims

1., based on a drop-down prompt system for input prefix, comprise client, at least one memory node and agent node:

Wherein, client, is suitable for receiving the request of search target keywords, and the request of described search target keywords is sent to agent node, comprise current input prefix in the request of described search target keywords;

Described agent node, comprising:

Inquiry forwarding module, is suitable for described search target keywords request forward to corresponding memory node;

Data forwarding module, the candidate item of the current input prefix of the correspondence being suitable for memory node to return is forwarded to client, and the candidate item of described correspondence current input prefix is the drop-down prompting result of recommending probability large;

Memory node, is suitable for based on the request of described search target keywords, obtain and the candidate item sending corresponding current input prefix to agent node;

Wherein, described memory node, comprising:

Recommend probability evaluation entity, be suitable for the weight according to each input prefix in the input prefix branch that is root with described current input prefix, and the weight of drop-down prompting result under each input prefix, calculate the probability that the drop-down prompting result under each input prefix is recommended;

Described recommendation probability evaluation entity, comprising:

First transition probability calculating sub module, the transition probability between being suitable for according to the weight calculation of described each input prefix input prefix.

2. system according to claim 1, is characterized in that, also comprises:

Meta data server, is suitable for safeguarding the overall routing table information pointing to memory node;

Described agent node, also comprises:

Routing table synchronization module, is suitable for the overall routing table information of meta data server to be synchronized to agent node;

Routing table update module, is suitable for upgrading overall routing table information from meta data server.

3. system according to claim 2, is characterized in that, described inquiry forwarding module inquires about corresponding memory node based on overall routing table information.

4. system according to claim 1, is characterized in that, described memory node, comprising:

Current input prefix determination module, is suitable for determining current input prefix;

Weight statistical module, is suitable for adding up with in the described current input prefix input prefix branch that is root, the weight of each input prefix, and the weight of drop-down prompting result under each input prefix;

Sequencing display module, is suitable for the probability arranged in sequence of described drop-down prompting result according to described recommendation, and selects to recommend the large drop-down prompting result of probability to be presented at client as the candidate item of current input prefix;