CN101661487B - Method and system for searching information items - Google Patents

Method and system for searching information items Download PDF

Info

Publication number
CN101661487B
CN101661487B CN2008102133341A CN200810213334A CN101661487B CN 101661487 B CN101661487 B CN 101661487B CN 2008102133341 A CN2008102133341 A CN 2008102133341A CN 200810213334 A CN200810213334 A CN 200810213334A CN 101661487 B CN101661487 B CN 101661487B
Authority
CN
China
Prior art keywords
user
evaluation
information
item
emotion
Prior art date
Application number
CN2008102133341A
Other languages
Chinese (zh)
Other versions
CN101661487A (en
Inventor
祝慧佳
蔡柯柯
郭宏蕾
苏中
Original Assignee
国际商业机器公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 国际商业机器公司 filed Critical 国际商业机器公司
Priority to CN2008102133341A priority Critical patent/CN101661487B/en
Publication of CN101661487A publication Critical patent/CN101661487A/en
Application granted granted Critical
Publication of CN101661487B publication Critical patent/CN101661487B/en

Links

Abstract

The invention provides a method and a system for searching information items. The method comprises the following steps: receiving the query for an object; based on the query, searching information items related to the object to acquire the information item set of the object, wherein each information item is related to a user; extracting the historical information item record of each related user in the information item set; based on the historical information item records of all users, calculating the effective weight of each user for the object; based on the effective weight, sorting all information items in the information item set of the object; and outputting the ordered information items as searching results.

Description

The method and system that item of information is searched for

Technical field

Present invention relates in general to search, more specifically, the present invention relates to assessment is searched for to object method and related system and computer program item of information.

Background technology

The continuous development of Along with computer technology and network technology, great deal of information is through the computer network transmission.Internet universal causes the growth by leaps and bounds of information sharing technology, and the internet also more and more is penetrated in people's the life, when people share information on the internet, also faces the own required information that how from so huge quantity of information, to retrieve effectively.

Present widely used search engine promptly is intended to assisting users to be retrieved from bulk information, so that the user can search and retrieve its required various information quickly and easily.Because the quantity and the numerous types of internet information content; Therefore the number as a result that search engine searched is quite a few; Useful is through Search Results being sorted, making that the most relevant, most important result comes the foremost, so that the user can obtain information more quickly.Therefore, improvement and optimization to various search engine techniques all are reacted directly in the ordering to Search Results.

The type of internet information content is abundant; Wherein along with the raising of web utilization rate and the quick expansion of ecommerce, increasing people searches information that is associated with various products (such as clothes, electronic product etc.) or the information that is associated with various services (service items such as for example food and drink, lodging, tourism) on web.These information generally include some and have used or bought evaluation and the suggestion of the user of these products or service to them, and these are estimated and suggestion in fact all is very useful for potential client, goods producer and online merchants.The information project that how these search is obtained or from corresponding data storage, directly obtain is a difficult problem of information retrieval field according to certain rank order, be shown to the user with more effective mode.Comparatively common method is that evaluation of user and suggestion were sorted according to the time at present, also is about to nearest evaluation of time and is placed on topmost, and the more past next time more early.But, thisly can't embody the importance of various evaluations, the information that also promptly can't help the user to obtain to expect effectively according to the method for time-sequencing.

Fig. 1 schematically shows in the prior art screenshot capture that Search Results that the user is estimated shows.As can be seen from Figure 1; Always having 473 users estimates; The star scoring that wherein specific products or service is provided according to each bar evaluation (comprise 1 star, 2 stars ..., 5 stars) shown intuitively that with histogram the user of each star scoring estimates number, and provided concrete evaluation number through the numeral in the bracket on the right side of histogram.In addition, giving average user estimates.The latter half at Fig. 1 has shown " the most useful user estimates ", and as can be seen from the figure, 759 philtrums have 716 people to think that its listed down evaluation of great use.

Above-mentioned prior art is to come grade assessment is carried out in various evaluations and suggestion through obtain feedback information from the user, thereby the demonstration of this " the most useful user estimates " is provided.Fig. 2 schematically shows the screenshot capture of gathering field feedback in the prior art.For example after the user reads the evaluation or suggestion that other users provide, can require the user that this evaluation or suggestion are given a mark or graded, thereby be that this is estimated or suggestion obtains a score value.In Fig. 2, click by the user through two buttons " Yes " and " No " simply and select this evaluation whether useful it.Based on statistics to these field feedbacks, can be to estimating or the suggestion ordering of giving a mark, with score value the highest come the foremost, what score value was low comes at the back.But there is following defective at least in said method: its needs are artificial participates in marking or grading, and its accuracy depends on the response rate of user feedback basically.And this, feedback answer person few network presence many for present reader is difficult to add up effectively.

Therefore; Need a kind ofly this type of information item to be searched for effectively, Search Results is sorted but not need the method for user feedback; Make item of information most important to the user, most worthy is come the foremost; Thereby help the more direct information that obtains product or service fast and effectively of user, also promptly just can obtain useful information as much as possible from a spot of item of information.

Summary of the invention

Therefore, in order to overcome the deficiency that exists in the prior art, the invention provides a kind of method and related system and computer program that item of information is searched for, it is suitable for method that the various assessments of product or service are searched for especially.

According to an aspect of the present invention, a kind of method that item of information is searched for is provided, comprises: receive inquiry object; Based on this inquiry, search for the item of information relevant to obtain the collection of information items of object with said object, wherein each item of information is associated with a user; Extract the historical information item records of each related user in this collection of information items; Based on all users' historical information item records, calculate the validity weight of each user to said object; Based on this validity weight, all items of information in the collection of information items of said object are sorted; And with the ordering item of information export as Search Results.

According to a further aspect in the invention, a kind of system that item of information is searched for is provided, comprises: receiving unit receives the inquiry to object; Search component based on this inquiry, is searched for the item of information relevant with said object to obtain the collection of information items of object, and wherein each item of information is associated with a user; Extraction assembly extracts the historical information item records of each related user in this collection of information items; The validity weight calculation component based on all users' historical information item records, is calculated the validity weight of each user to said object; Sequencing assembly based on this validity weight, sorts to all items of information in the collection of information items of said object; And output precision, the item of information of ordering is exported as Search Results.

According to a further aspect in the invention, a kind of computer program that is used to carry out computer program code according to the method for the invention that comprises also is provided.

Utilize the present invention, the user is to product or service when searching for, and its most worthy, item of information that reliability is the highest come usually the more forward position of information project tabulation.

Further, search ordering method of the present invention can use separately to the item of information in database or the data storage, also can be used in combination with any existing search engine, optimizes the demonstration to its Search Results.

Description of drawings

With reference to after the detailed description below in conjunction with accompanying drawing, it is more obvious that characteristic of the present invention, advantage and others will become, wherein in the accompanying drawings:

Fig. 1 shows in the prior art screenshot capture that Search Results that the user is estimated shows;

Fig. 2 shows the screenshot capture of gathering field feedback in the prior art;

Fig. 3 shows process flow diagram according to the method for the invention;

Fig. 4 shows the Organization Chart according to system of the present invention; And

Fig. 5 shows and can realize computer system of the present invention.

Note that if exist, identical reference marker is represented identical parts in whole accompanying drawings.

Embodiment

In following detailed description,, many specific details have been described through the mode of example for the ease of complete understanding the present invention.But those skilled in the art can be very clear, and the present invention also can not need these details just can realize.In addition, in order more clearly to explain the present invention, in some example, known method, processing, element and circuit only are to have carried out describing synoptically, and do not specify.Below in conjunction with accompanying drawing the present invention is explained in more detail and explains.Should be appreciated that accompanying drawing of the present invention and embodiment only are used for exemplary effect, be not to be used to limit protection scope of the present invention.

The present invention relates to the item of information of various objects is searched for and sorted.Wherein, " object " can comprise various tangible or invisible product, product feature and/or services, includes but not limited to, for example clothes, digital product, hotel, food and drink, tourism or the like.Can provide relevant evaluation to all these objects, the item of information of these objects for example can be to use or buy evaluation or the suggestion of the user of this series products or service to them.

Below will describe method and system of the present invention through instantiation, wherein in this example, we are with to the evaluation of various products or service or the suggestion example as item of information.But, it will be understood by those skilled in the art that to the invention is not restricted to this, but can be applied in the sort method to the various items of information that obtain through the search of any approach.

In existing various electronic business transaction website; For a product or service; Usually can provide and use/bought this product or enjoyed the evaluation information of the user of this service to this product or service, these information can help the user to judge whether to be fit to select this product or service.But, along with the growth of evaluation information quantity, especially integrate under the situation of evaluation information of separate sources at need, can reach hundreds and thousands of for the quantity of the evaluation information of a product or service.The evaluation information of quantity is to be difficult to read one by one for the user like this.For this reason, the user hopes to obtain fast and effectively that reference value and objective appraisal information are arranged most.So; It is from said object being had than the use experience of horn of plenty or the evaluation information that the user delivered of experience that reference value and effectively evaluating are arranged for the user most, rather than professional writer or lack the evaluation information that the user of use experience delivers.And the user can be through the evaluation information acquisition of analysis user history to this class object with the evaluation objectivity to the use experience of said object.

Hence one can see that, and estimator's history evaluation is very useful for the evaluation criteria of analyzing this estimator.Can analyze through history evaluation, thereby the assessment as far as possible accurately to its evaluation that is directed against special object that provides is provided the specific user.

When assay person's history is estimated; We consider to overcome following problem; How to judge promptly whether the evaluation that the estimator provides is objective; If an estimator always is positive suggestion or always is negative suggestion then whether this estimator's suggestion valuable? Be and how to discern this estimator the writer that businessman employs?

To the problems referred to above, we are that the user of the item of information that provides various objects (perhaps estimating) distributes a validity weight, the validity of the item of information of this object of reflection in this validity weight.The user is sorted to the evaluation of the object validity weight according to the user, thereby the item of information of most worthy is placed on the foremost, make the user can fast and effeciently obtain useful information.

Because the object coverage rate estimated is extensive, therefore need classify to the object of being estimated so as more in an organized way, the validity weight of distributing user more reasonably.In existing shopping website or online shopping mall, there has been classification usually to product or service, for example, can be divided into dress ornament, footwear bag, digital product, books, phonotapes and videotapes, household articles, food, hotel or the like classification roughly.The item of information of various products or service is associated with regard to its corresponding classification.The division of above-mentioned classification can reasonably distribute, be provided with flexibly according to the quantity of product that is provided or service, type etc.Above-mentioned classification can be pre-set, also can after adjust, for example when the model of certain series products, brand etc. increase, can such be segmented again, perhaps when the product number tails off, merge some classification.

In order to solve the problem that possibly exist in the above-mentioned reality, comprise three factors at least in the estimator's that the present invention proposes the validity weight: the object type number that this estimator estimated accounts for the ratio of total object type number; This estimator is inclined to consistent degree to the evaluation of each object type of commenting on the emotion of the overall evaluation; This estimator's evaluation is shared proportion in all are estimated.

Estimator's validity weight is the funtcional relationship of these three factors.In one embodiment, (Ri is Oj) by classification ratio R to the validity weight A of special object classification Oj for a certain estimator Ri Category, emotion is inclined to consistent degree Con SentiWith evaluation proportion R CommentThe company of these three factors takes advantage of expression, also promptly:

A(Ri,Oj)=R category*Con senti*R comment(1)

It will be understood by those skilled in the art that the funtcional relationship that can also use other representes this validity weight, such as linear combination of three factors etc.

Classification ratio R CategoryNumber shared ratio in total object type number of representing the object type that specific estimator Ri was estimated.From its definition, can obtain its computing method intuitively, that is:

R CategoryThe number (2) of the object type of the number of the object type that=Ri estimated/total

Classification ratio R CategoryThe writer that the factor can mask businessman is effectively employed.This be because; The writer that common businessman is employed often only estimates to object (product or service) a kind of or a few classification; Because the number of its object type of estimating is less; Therefore, ratio occupied in all object type is also less, thereby can be reflected in this estimator's the validity weight through the classification scale factor.

Emotion is inclined to consistent degree Con SentiRepresent the emotion tendency consistance of specific estimator to the evaluation and the overall evaluation of each object type of commenting on.When the assay person is inclined to the emotion of each evaluation object classification, uses natural language processing and analyze, and it is expressed as emotion tendency vector S entiV (Ri).Along with computing machine and broad application of Internet, the accessible natural language text quantity of computing machine unprecedentedly increases, and natural language processing technique plays an important role in application such as the text mining that is directed against magnanimity information, information extraction, man-machine interaction.Natural language processing is started with from two aspects of syntax and semantics, from text, extracts the analysis that content corresponding is carried out the meaning of a word and sentence justice, excavates the expressed suggestion of estimator.Alternatively, can give certain numerical value with estimator's suggestion, for example from-5 to 5 numerical value.Clearly, it will be understood by those skilled in the art that the numerical value that also can get other scopes representes.Wherein each element of emotion tendency vector is the evaluation suggestion numerical value of this estimator Ri to each object type.

Maybe be from malevolence or irresponsiblely make evaluation to the above-mentioned estimator who has, provide positive or negative evaluation all the time, this evaluation is not worth basically.Therefore, through specific estimator's the evaluation and the emotion tendency of the overall evaluation are compared, can suppress above-mentioned phenomenon effectively.The emotion tendency of the overall evaluation is also represented by emotion tendency vector S entiV (all).SentiV (all) can be a statistical value, for example, gives up extremum mean value afterwards, perhaps only extracts the sample average afterwards of some.In one embodiment, the emotion of overall evaluation tendency vector S entiV (all) is represented by the average of all estimators' emotion tendency vector.

In one embodiment, the emotion of estimator Ri is inclined to consistent degree Con SentiThe emotion that is calculated as this estimator Ri is inclined to the inner product of the emotion tendency vector of the vector and the overall evaluation, promptly

Con senti=SentiV(Ri)·SentiV(all)(3)

Clearly, if estimator Ri is consistent to the suggestion of each object type with most of estimator's suggestion, then emotion is inclined to consistent degree Con SentiValue also higher.If for each object type; The suggestion of estimator Ri just is being always (score value is very high) or is being always negative (score value is very low), then compares in the emotion tendency with the overall evaluation, also promptly calculates after the inner product; Its numerical value is also little, thereby can be inclined to consistent degree Con through emotion SentiThe factor is reflected in the validity weight of this estimator Ri.

Estimate proportion R CommentEvaluation shared proportion in all are estimated of representing specific estimator.In one embodiment, estimate proportion and be made up of two parts, estimator Ri accounts for the proportion of this estimator to the evaluation number of all object type to the evaluation number of object type under the said evaluation object; And this estimator accounts for the proportion of all estimators to the evaluation number of object type under the said evaluation object to the evaluation number of object type under the said evaluation object.In one embodiment, can calculate evaluation proportion R through following formula Comment:

Wherein, coefficient lambda 1And λ 2Be to be used for these two parts of balance at evaluation proportion R CommentIn shared proportion, it can rule of thumb adjust λ 1And λ 2Value.Can find out from above-mentioned formula; With himself the evaluation number of all object type is compared and the evaluation number of Oj classification is compared through estimator Ri is done the evaluation number of appearing to the Oj classification with all estimators; Can judge that whether this estimator Ri lays particular emphasis on certain object type, that is to say it is not the authoritative sources in this object type field.At the evaluation proportion R of Ri to the Oj object type CommentWhen high, the evaluation of Oj is had more value, be reflected in that it can increase the numerical value of validity weight on the validity weight with regard to meaning Ri.

The validity weight relevant with item of information of embodiment of the present invention in a kind of the application described hereinbefore; Based on this validity weight; Can the item of information that search be sorted, thereby the high item of information of validity weight comes the front, the item of information that the validity weight is low comes the back; Make the user only need browse a spot of information, just can obtain desired Useful Information fast.

With reference to figure 3 method flow of searching for according to the item of information to object of embodiment of the present invention is described below.In this flow process, still with to describing such as the evaluation of various products or service object or suggestion example as item of information.

As shown in Figure 3, in step S300, begin this treatment scheme.

In step S302, receive the inquiry of user to evaluation object Oj.

In step S304; Based on user's inquiry, search for the item of information relevant to obtain the collection of information items of this evaluation object with evaluation object Oj, wherein each item of information is associated with a user; Also be that each item of information is generated by user associated therewith, this user is also referred to as the estimator.In one embodiment; For example; When being used in combination, according to the evaluation object inquiry of user's input among the step S302, the item of information that from each database, is associated with this evaluation object based on keyword search with search engine; Also promptly to the evaluation or the suggestion of this evaluation object, thereby obtain the collection of information items of object Oj.From these evaluations or suggestion, can obtain to generate user's's (being the estimator) of this evaluation or suggestion ID, thereby form estimator's set of this evaluation object.In another embodiment; When various objects have carried out the branch time-like according to the mode that the front is described; The evaluation of each object can be associated with the classification under it in data storage; Also promptly index is carried out in the object evaluation, thereby can search out collection of information items and estimator's set of institute's query object fast through classification.

In step S306, the historical information item records of each related user in the set of information extraction item is also promptly extracted each estimator's in estimator's set historical evaluation record.For example, according to estimator's ID, all once gave the evaluation record of appearing to retrieve this estimator.

In following step, with calculating the validity weight that each estimator Ri is directed against the evaluation object Oj that is inquired about.

In step S308, the classification of all objects that identification and evaluation person Ri was estimated, the number of objects of statistics classification.This result can be applied in the classification scale factor in the calculating validity weight subsequently.

In step S310, calculate the validity weight that estimator Ri is directed against the evaluation object Oj that is inquired about., preceding text can know that the validity weight of embodiment of the present invention comprises three factors at least from describing: classification ratio R Category, the object type number that estimator Ri estimated accounts for the ratio of total object type number; Emotion is inclined to consistent degree Con Senti, estimator Ri is inclined to consistent degree to the evaluation of each object type of commenting on the emotion of the overall evaluation; Estimate proportion R Comment, the evaluation of estimator Ri is shared proportion in all are estimated.In this step, according to the formula that the front provides, calculate this three factors respectively, for example three factors are connected then and multiply by the value that obtains the validity weight.

Wherein, classification ratio R CategoryBe calculated as: the number that uses the object type that Ri estimated in step S308, obtain is divided by total classification number, and this total classification number can be in that preamble is described is definite when all objects are classified.

Emotion is inclined to consistent degree Con SentiCalculating relate to: at first use natural language processing technique assay person to the emotion in the evaluation of each object type tendency, for example give the degree that certain numerical value is represented its emotion tendency.In one embodiment, positive evaluation Example is as with 1 to 5 numeric representation, and negative evaluation Example is as with-5 to-1 numeric representation, 0 expression neutrality.Then, for each estimator Ri, set up the emotion tendency vector S entiV (Ri) of this estimator Ri.Generally speaking, the dimension of the emotion of estimator Ri tendency vector is the number of the object type estimated of Ri, and each element is the evaluation suggestion numerical value of this estimator Ri to corresponding object type.Because each estimator not necessarily can estimate all object type, make like this that therefore the dimension of emotion tendency vector separately is not corresponding.For unified these vectors and convenience of calculation afterwards in form, it is the number of total object type that the dimension of all emotion tendency vectors all unify, does not make the object type of any evaluation for estimator Ri, and its elements corresponding value is 0.Then, after having obtained all estimators' emotion tendency vector, in one embodiment, can ask on average, to obtain the emotion tendency vector S entiV (all) of the overall evaluation all these vectors.Clearly, it will be appreciated by those skilled in the art that the account form that can also adopt other finds the solution the emotion tendency vector of the overall evaluation according to all estimators' emotion tendency vector.At last, the emotion tendency vector S entiV (Ri) of estimator Ri and the emotion tendency vector S entiV (all) of the overall evaluation are carried out inner product, be inclined to consistent degree Consenti with the emotion that obtains estimator Ri.

Estimate the calculating R of proportion CommentComprise: calculating estimator Ri accounts for the proportion of this estimator Ri to the evaluation number of all object type to the evaluation number of the classification of evaluation object Oj; And calculate estimator Ri the evaluation number of the classification of evaluation object Oj is accounted for the proportion of all estimators to the evaluation number of the classification of evaluation object Oj.Wherein, estimator Ri is exactly the number of all evaluations of the given mistake of this estimator Ri to the evaluation number of all object type basically.All estimators to the evaluation number of the classification of evaluation object Oj be all estimators in estimator's set of in step S304, obtaining above-mentioned to the evaluation number of the classification of evaluation object Oj, it is that all of classification of object Oj are estimated numbers basically.Then, calculate evaluation proportion R according to aforementioned formula (4) Comment, experience factor λ 1And λ 2Can these two parts of balance result estimate proportion R CommentIn shared proportion.

After calculating above-mentioned three factors, just can obtain the validity weight of estimator Ri to object Oj.

Then, in step S312, the validity weight that judges whether all estimators has all been calculated and has been finished.

If judged result is " denying ", then handle and turn back to step S308, continue next estimator is calculated, and continue to carry out later step.

If judged result is " being ", then handles and advance to step S314.

In step S314, all validity weights of estimating based on each estimator of selected evaluation object Oj to be sorted, the evaluation that wherein the validity weight is high has more forward position.Through considering the validity weight of estimator, can obtain the effect that better sorts to specific evaluation object.

At last, in step S316, will pass through the item of information of ordering and export as Search Results.In one embodiment, for example can on display screen, Search Results be presented to the user, wherein for example can only show the preceding 10-50 bar item of information after sorting, so that the user browses.It will be understood by those skilled in the art that and also can adopt various graphical informations to present to the user.

Alternatively, the ordering to Search Results of the present invention can also be further considered other factors except considering estimator's validity weight, for example provides the time of evaluation.For the object (product or service) of some classification, the evaluation that early provides possibly be worth little for current evaluation object, so after its position when ordering leans on.

Fig. 4 schematically shows the system schematic block diagram of realization according to the searching method of one embodiment of the present invention.Wherein, Reference number 400 expressions comprise in this search system 400: receiving unit 401, search component 402, extraction assembly 404, validity weight calculation component 406, sequencing assembly 408 and output precision 409 according to the search system of embodiment of the present invention.Wherein, in Fig. 4, also show data repository 410, this data repository 410 can be included among the ordering system 400, also can be used as its outside individual components, even this data repository 410 can also be distributed in the computer network with distributed.

As shown in Figure 4, receiving unit 401 for example receives the inquiry to object from the user.Search component 402 is for example searched for the item of information relevant with the object of being inquired about based on the inquiry that receives from data repository 410, the collection of information items of query object to obtain, and wherein each item of information is associated with a user.Search component 402 is all associated user of search from the collection of information items of obtaining also, form user's set.Extraction assembly 404 extracts historical information item records separately for each user in this user's set then, for example also from data repository 410, obtains these historical information item records.Then, validity weight calculation component 406 is calculated the validity weight of each user to special object.Then, based on the validity weight of being calculated, sorted by all items of information in the collection of information items of 408 pairs of these special objects of sequencing assembly, wherein, the item of information that the validity weight is high is arranged in more forward position, thus the ordering effect that is improved.At last, will pass through the item of information of ordering and export by output precision 409, for example present to the user through display screen as Search Results.

In a kind of application of the present invention, for example, the item of information of object is evaluation or the suggestion of user to this object.In this application, validity weight calculation component 406 can comprise following sub-component at least: classification ratio computation module 412, emotion are inclined to consistent degree computation module 414, are estimated proportion computation module 416, function assembly 418 and recognizer component 420.

Wherein, classification ratio computation module 412, emotion are inclined to consistent degree computation module 414 and are estimated proportion computation module 416 and calculate the corresponding factor according to method described above respectively.Be inclined in emotion and also comprise natural language processing assembly (not shown) in the consistent degree computation module 414, this natural language processing assembly can be analyzed the emotion tendency of expressing in each item of information according to syntax and semantics, and it is expressed as emotion tendency vector.

Classification under recognizer component 420 identifying objects; Also promptly discern the classification of each object in the historical information item records of extracting by extraction assembly 404; Thereby count the number of relevant object type or the evaluation number of related category, to be used for follow-up use or calculating.For example, in classification ratio computation module 412, be used to calculate the ratio that object type number that estimator Ri estimated accounts for total object type number; Be inclined in the consistent degree computation module 414 in emotion, the object type that can use recognizer component 420 to discern in each item of information is inclined to vector to assist setting up emotion; And in estimating proportion computation module 416, be used to calculate estimator Ri to evaluation number of the classification of estimator's object Oj or the like.

Function assembly 418 receives the result of calculation that classification ratio computation module 412, emotion are inclined to consistent degree computation module 414 and estimate proportion computation module 416, calculates the validity weight according to funtcional relationship.In one embodiment, the function assembly for example can be a multiplier, and its three results that will receive (being classification ratio, emotion tendency degree and evaluation proportion) multiply each other, to obtain the validity weight.In optional embodiment, the function assembly can also be totalizer or totalizer and the combining or the funtcional relationship computation module of any appropriate of multiplier.

In optional embodiment, sequencing assembly 408 can also have other inputs, time of generating of each item of information for example, thus sequencing assembly 408 can further sort to item of information according to the rise time.

Fig. 4 only shows the example that can realize a kind of ordering system of the present invention.It will be understood by those skilled in the art that in the concrete realization of each components/modules, when especially realizing each functions of modules, can have the plurality of optional scheme through software.For example, can with shown in search component 402 merge in the same module with extraction assembly 404.

Below, will describe with reference to figure 5 and can realize computer system of the present invention.Fig. 5 has schematically shown the block diagram that can realize computer system according to the embodiment of the present invention.

Computer system shown in Fig. 5 comprises CPU (CPU) 501, RAM (RAS) 502, ROM (ROM (read-only memory)) 503, system bus 504, hard disk controller 505, KBC 506, serial interface controller 507, parallel interface controller 508, display controller 509, hard disk 510, keyboard 511, serial external unit 512, parallel external unit 513 and display 514.In these parts, what link to each other with system bus 504 has CPU501, RAM502, ROM503, hard disk controller 505, KBC 506, serial interface controller 507, parallel interface controller 508 and a display controller 509.Hard disk 510 links to each other with hard disk controller 505; Keyboard 511 links to each other with KBC 506; Serial external unit 512 links to each other with serial interface controller 507, and parallel external unit 513 links to each other with parallel interface controller 508, and display 514 links to each other with display controller 509.

Each functions of components all is well-known in the present technique field among Fig. 5, and shore and structure shown in Figure 5 also are conventional.The described block diagram of Fig. 5 illustrates just to the purpose of example, is not to be limitation of the present invention.In some cases, can add or reduce wherein some parts as required.

Search of the present invention and Search Results is carried out sort method can use separately to the item of information in database or the data storage also can be used in combination with any existing search engine, optimizes the demonstration to its Search Results.Can know that from the description of preceding text the process that item of information is searched for can realize with any known algorithm, process, mode, and should realization itself not belong to scope of the present invention.

In addition; It will be appreciated by those skilled in the art that; Although various aspects of the present invention can be used as block diagram, process flow diagram or use other diagram to represent to explain and describe; But be appreciated that these modules described here, assembly, equipment, system, technology or method can realize to make up as hardware, software, firmware, special circuit or logic, common hardware or the controller of limiting examples or other computing equipment or its.

Although instruction of the present invention is in the context of practical implementation, to describe, it will be apparent to one skilled in the art that under the situation that does not break away from spirit of the present invention, can make amendment and change each embodiment of the present invention.Description in this instructions is only used for illustrative, and should not be considered to restrictive.Scope of the present invention only receives the restriction of appended claims.

Claims (20)

1. method that item of information is searched for comprises:
Reception is to the inquiry of object;
Based on this inquiry, search for the item of information relevant to obtain the collection of information items of object with said object, wherein each item of information is associated with a user;
Extract the historical information item records of each related user in this collection of information items;
Based on all users' historical information item records, calculate the validity weight of each user to said object;
Based on this validity weight, all items of information in the collection of information items of said object are sorted; And
The item of information of ordering is exported as Search Results,
Wherein calculating each user comprises to the validity weight of said object:
Calculate the ratio that object type number that this user estimated accounts for total object type number;
Calculate this user the evaluation of each object type of estimating is inclined to consistent degree with the emotion of the overall evaluation;
Calculate this evaluation of user shared proportion in all are estimated.
2. method according to claim 1, wherein said item of information are the evaluation of user to said object.
3. method according to claim 1 and 2 also comprises:
Discern the affiliated object type of said object.
4. method according to claim 2, wherein calculate each user and also comprise to the validity weight of said object:
Classification ratio, emotion are inclined to consistent degree and are estimated proportion and multiply by mutually and obtain said validity weight.
5. method according to claim 4 is wherein used natural language processing analysis user from the user estimates the emotion of each object type is inclined to, and it is expressed as emotion tendency vector.
6. method according to claim 5 is wherein calculated this user and the evaluation of each object type of estimating is inclined to consistent degree with the emotion of the overall evaluation is comprised: this user's emotion tendency vector is inclined to vector with the emotion of the overall evaluation carries out inner product.
7. method according to claim 6, the emotion tendency vector of the wherein said overall evaluation is represented by the average of all users' emotion tendency vector.
8. according to the arbitrary described method of claim 4-7, wherein calculate said evaluation proportion according to following factor: this user accounts for the proportion of this user to the evaluation number of all object type to the evaluation number of object type under the said object; And this user accounts for the proportion of all users to the evaluation number of object type under the said object to the evaluation number of object type under the said object.
9. method according to claim 1, wherein that the validity weight is high item of information comes the front.
10. method according to claim 1, wherein search step comprises the search item of information relevant with said object from least one data repository.
11. the system that item of information is sorted comprises:
Receiving unit receives the inquiry to object;
Search component based on this inquiry, is searched for the item of information relevant with said object to obtain the collection of information items of object, and wherein each item of information is associated with a user;
Extraction assembly extracts the historical information item records of each related user in this collection of information items;
The validity weight calculation component based on all users' historical information item records, is calculated the validity weight of each user to said object;
Sequencing assembly based on this validity weight, sorts to all items of information in the collection of information items of said object; And
Output precision is exported the item of information of ordering as Search Results,
Wherein the validity weight calculation component comprises:
Classification ratio computation module calculates the object type number that this user estimated and accounts for total object class ratio of number in addition;
Emotion is inclined to consistent degree computation module, calculates this user the evaluation of each object type of estimating is inclined to consistent degree with the emotion of the overall evaluation;
Estimate the proportion computation module, calculate this evaluation of user shared proportion in all are estimated.
12. system according to claim 11, wherein said item of information is the evaluation of user to said object.
13., also comprise according to claim 11 or 12 described systems:
Recognizer component is used to discern the affiliated object type of said object.
14. system according to claim 13, wherein the validity weight calculation component further comprises:
Multiplier is inclined to said classification ratio, emotion consistent degree and estimates proportion and multiply by mutually and obtain said validity weight.
15. system according to claim 14 wherein is inclined in the consistent degree computation module in emotion, comprises the natural language processing assembly, is used for estimating the emotion tendency of analysis user to each evaluation object classification from the user, and it is expressed as emotion tendency vector.
16. system according to claim 15, wherein emotion is inclined to consistent degree computation module and is comprised this user's the emotion tendency vector and the emotion of the overall evaluation are inclined to the assembly that vector carries out inner product.
17. system according to claim 16, the emotion tendency vector of the wherein said overall evaluation is represented by the average of all users' emotion tendency vector.
18. according to the arbitrary described system of claim 14-17, wherein estimate the proportion computation module and calculate said evaluation proportion according to following factor: this user accounts for the proportion of this user to the evaluation number of all object type to the evaluation number of object type under the said object; And this user accounts for the proportion of all users to the evaluation number of object type under the said object to the evaluation number of object type under the said object.
19. system according to claim 11, the item of information that wherein sequencing assembly is high with the validity weight comes the front.
20. system according to claim 11, wherein search component is searched for the item of information relevant with said object from least one data repository.
CN2008102133341A 2008-08-27 2008-08-27 Method and system for searching information items CN101661487B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2008102133341A CN101661487B (en) 2008-08-27 2008-08-27 Method and system for searching information items

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2008102133341A CN101661487B (en) 2008-08-27 2008-08-27 Method and system for searching information items

Publications (2)

Publication Number Publication Date
CN101661487A CN101661487A (en) 2010-03-03
CN101661487B true CN101661487B (en) 2012-08-08

Family

ID=41789513

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2008102133341A CN101661487B (en) 2008-08-27 2008-08-27 Method and system for searching information items

Country Status (1)

Country Link
CN (1) CN101661487B (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102314443B (en) * 2010-06-30 2016-02-17 深圳市世纪光速信息技术有限公司 The modification method of search engine and system
US9183573B2 (en) * 2011-06-03 2015-11-10 Facebook, Inc. Überfeed
JP5476339B2 (en) * 2011-06-22 2014-04-23 楽天株式会社 Information processing apparatus, information processing method, information processing program, and recording medium on which information processing program is recorded
CN102982035B (en) * 2011-09-05 2015-10-07 腾讯科技(深圳)有限公司 A kind of search ordering method of community users and system
US20130179230A1 (en) * 2012-01-05 2013-07-11 International Business Machines Corporation Organizational agility improvement and prioritization across multiple computing domains
CN103577988B (en) * 2012-07-24 2017-08-04 阿里巴巴集团控股有限公司 A kind of method and apparatus for recognizing specific user
CN102970352A (en) * 2012-11-08 2013-03-13 北京小米科技有限责任公司 Method and device for providing information
CN103853948B (en) * 2012-11-28 2018-01-19 阿里巴巴集团控股有限公司 The identification of user identity, the filtering of information and searching method and server
CN105051776A (en) * 2013-03-25 2015-11-11 索尼公司 Information processing device, information processing system, and information processing method
CN104281619A (en) * 2013-07-11 2015-01-14 鸿富锦精密工业(深圳)有限公司 System and method for ordering search results
CN104516897B (en) * 2013-09-29 2018-03-02 国际商业机器公司 A kind of method and apparatus being ranked up for application
CN103646092B (en) * 2013-12-18 2017-07-04 孙燕群 Based on the method for sequencing search engines that user participates in
CN104866474B (en) * 2014-02-20 2018-10-09 阿里巴巴集团控股有限公司 Individuation data searching method and device
CN106776708A (en) * 2016-11-17 2017-05-31 中国农业银行股份有限公司 A kind of querying method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1559044A (en) * 2001-11-26 2004-12-29 富士通株式会社 Content information analyzing method and apparatus
CN1605201A (en) * 2001-12-13 2005-04-06 皇家飞利浦电子股份有限公司 Recommending media content on a media system
CN1818908A (en) * 2006-03-16 2006-08-16 董崇军 Feedbakc information use of searcher in search engine

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1559044A (en) * 2001-11-26 2004-12-29 富士通株式会社 Content information analyzing method and apparatus
CN1605201A (en) * 2001-12-13 2005-04-06 皇家飞利浦电子股份有限公司 Recommending media content on a media system
CN1818908A (en) * 2006-03-16 2006-08-16 董崇军 Feedbakc information use of searcher in search engine

Also Published As

Publication number Publication date
CN101661487A (en) 2010-03-03

Similar Documents

Publication Publication Date Title
JP6152173B2 (en) Ranking product search results
Tan et al. Rating-boosted latent topics: Understanding users and items with ratings and reviews.
Matta et al. Bitcoin Spread Prediction Using Social and Web Search Media.
CN104951428B (en) User's intension recognizing method and device
CN103324657B (en) Extended using the multi-dimensional query of semantic and service condition statistics
US20150058331A1 (en) Search result ranking using machine learning
US8756095B2 (en) E-commerce consumer-based behavioral target marketing reports
CN102866990B (en) A kind of theme dialogue method and device
JP5596152B2 (en) Information matching method and system on electronic commerce website
CN102279851B (en) Intelligent navigation method, device and system
CN103678335B (en) The method of method, apparatus and the commodity navigation of commodity sign label
CN101551806B (en) Personalized website navigation method and system
US8219456B2 (en) Method, device, and system for analyzing and ranking web-accessible data targets
CN103430202B (en) Searching system adopting multilevel silhouettes
JP5368665B2 (en) Expert database forwarded back to link weighted association rules
TWI539305B (en) Personalized information push method and device
CN105794154A (en) System and method for analyzing and transmitting social communication data
CN103729359B (en) A kind of method and system recommending search word
US20100313141A1 (en) System and Method for Learning User Genres and Styles and for Matching Products to User Preferences
CN105247507B (en) Method, system and storage medium for the influence power score for determining brand
JP6134444B2 (en) Method and system for recommending information
US8738436B2 (en) Click through rate prediction system and method
CN103154991B (en) Credit risk is gathered
CN104166668A (en) News recommendation system and method based on FOLFM model
JP6247292B2 (en) Query expansion

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
C14 Grant of patent or utility model
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20120808

Termination date: 20150827

EXPY Termination of patent right or utility model