CN105302898A - Click model-based searching and ranking method and device - Google Patents
Click model-based searching and ranking method and device Download PDFInfo
- Publication number
- CN105302898A CN105302898A CN201510697625.2A CN201510697625A CN105302898A CN 105302898 A CN105302898 A CN 105302898A CN 201510697625 A CN201510697625 A CN 201510697625A CN 105302898 A CN105302898 A CN 105302898A
- Authority
- CN
- China
- Prior art keywords
- result
- score value
- sequence
- order
- result items
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a click model-based searching and ranking method. The method comprises the following steps: obtaining a first result sequence in regard to query, wherein each result item in the first result sequence has a first score, all the result items are ranked from high to low according to the first score, and the first score is calculated according to at least one predetermined characteristic; determining the second score of each result item in the first result sequence by utilizing a click model; and selecting the result items with the second score not equal to 0 in the first result sequence, and adjusting the sequence of the selected result items on the basis of the second scores so as to obtain a second result sequence. The invention furthermore discloses a click model-based searching and ranking device.
Description
Technical field
The present invention relates to search engine, especially a kind of search ordering method based on click model and device.
Background technology
Internet, for the effect of economic society, from abatement information asymmetry, is extending to and is driving large data to produce value.In this process, search is people from the important means of internet obtaining information and data all the time, therefore becomes the important entrance of internet.Promoting the retrieval relevance of search engine, is an important research direction of information retrieval field.In recent years, the click model in information retrieval field achieves tremendous development.So-called click model, utilize the click data of user when using search engine exactly, calculating Search Results and user inquire about the mathematical model of correlativity.It makes search system when ranking results, obtains the ability of user's click information before using, thus the high result of correlativity can be allowed to arrange more forward.
Although utilize click model to carry out click tune power have good dependence query result, click to adjust and weigh and the weight order be not suitable in decision search separately.First, the high and correlativity of clicking rate well and be not exclusively the same thing; Secondly, the essence sorted in theory is its correlativity of information computing that reasonable employment result comprises, and click has openness feature, namely the sub-fraction that the inquiry that user clicked result is just all inquired about, and the result that user clicked in an inquiry is the sub-fraction of this Query Result, therefore it can be limited for the quantity of information calculating correlativity contribution.To sum up, clicking tune power should use together with the existing further feature of search engine, namely it must be dissolved in existing sequence framework.But click tune power and participate in sequence, often cause " positive feedback " problem, namely the clicked result adjusting power to be discharged to above, often obtains higher click, thus before being more easily discharged to.
Therefore, how adjusting power reasonably to incorporate in existing sequence framework click, and do not cause harmful effect, is the problem to be solved in the present invention.
Summary of the invention
For this reason, the invention provides a kind of search ordering method based on click model and device, to try hard to solve or at least alleviate at least one problem existed above.
According to an aspect of the present invention, provide a kind of search ordering method based on click model, the method comprising the steps of: obtain the first result sequence about inquiry, in first result sequence, each result items has the first score value, and all result items are according to the order sequence from high to low of the first score value, the first score value calculates according at least one predetermined characteristic; Click model is utilized to determine the second score value of each result items in the first result sequence; Choosing the second score value in the first result sequence is not the result items of 0, based on the order of the result items selected by the second score value adjustment, obtains the second result sequence.
Alternatively, according to of the present invention based in the search ordering method of click model, it is one or more that predetermined characteristic comprises in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity.
Alternatively, according to of the present invention based in the search ordering method of click model, click model be series connection click model.
Alternatively, according to of the present invention based in the search ordering method of click model, the step of order based on the result items selected by the second score value adjustment comprises: the result items the second score value being greater than to threshold value, according to the order sequence from high in the end of the second score value; Second score value is not more than to the result items of threshold value, maintenance order is constant, and comes after the second score value is greater than all result items of threshold value.
Alternatively, according to of the present invention based in the search ordering method of click model, threshold value be series connection click model codomain average.
Alternatively, according to of the present invention based in the search ordering method of click model, also comprise step: by the first score value of the result items that position in the second result sequence changes, be adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items.
According to a further aspect in the invention, provide a kind of searching order device based on click model, this device comprises: acquiring unit, is suitable for obtaining the first result sequence about inquiry, and wherein in the first result sequence, each result items has the first score value; Computing unit, is suitable for calculating the first score value according at least one predetermined characteristic, being also suitable for utilizing click model to determine the second score value of each result items in described first result sequence; And sequencing unit, being suitable for all result items in the first result sequence to sort according to the first score value order from high to low, be also suitable for choosing the second score value in the first result sequence is not the result items of 0, based on the order of the result items selected by the second score value adjustment, to obtain the second result sequence.
Alternatively, according to of the present invention based in the searching order device of click model, it is one or more that predetermined characteristic comprises in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity.
Alternatively, according to of the present invention based in the searching order device of click model, click model be series connection click model.
Alternatively, based in the searching order device of click model, sequencing unit also comprises judgment sub-unit according to of the present invention, the second score value be suitable for selected by judging is not that in the result items of 0, whether the second score value is greater than threshold value; Sequencing unit is also suitable for the result items the second score value being greater than to threshold value, according to the order sequence from high in the end of the second score value; And the second score value being not more than to the result items of threshold value, maintenance order is constant, and comes after the second score value is greater than all result items of threshold value.
Alternatively, according to of the present invention based in the searching order device of click model, threshold value be series connection click model codomain average.
Alternatively, according to of the present invention based in the searching order device of click model, sequencing unit is also suitable for the first score value of the result items changed position in the second result sequence, is adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items.
According to another aspect of the present invention, provide a kind of information search engine system, comprising: information bank, be suitable for storing information to be put; As above based on the searching order device of click model, be suitable for sorting to inquiring about the result sequence obtained; And information display device, be suitable for showing Query Result in order.According to the searching order scheme based on click model of the present invention, on the basis of old collating sequence, in conjunction with the model feature of series connection click model, reasonably power is adjusted to be dissolved in the framework of searching order click, change position and the weights of result items, make the final DISPLAY ORDER of Query Result, not only reflect the positive effect of click model to correlativity, also avoid positive feedback shortcoming.
In addition, according to the solution of the present invention, maintain the second result sequence still according to the character of old sequence, make all multioperations based on this character in search engine, such as this sequence and the result sequence of other parallel search engines are combined together minor sort again, still effective after use this method.
Accompanying drawing explanation
In order to realize above-mentioned and relevant object; combine description below and accompanying drawing herein to describe some illustrative aspect; these aspects indicate the various modes can putting into practice principle disclosed herein, and all aspects and equivalent aspect thereof are intended to fall in the scope of theme required for protection.Read detailed description below in conjunction with the drawings, above-mentioned and other object of the present disclosure, Characteristics and advantages will become more obvious.Throughout the disclosure, identical Reference numeral is often referred to for identical parts or element.
Fig. 1 shows the exemplary environments 100 that search engine according to an embodiment of the invention runs wherein;
Fig. 2 shows according to an embodiment of the invention based on the process flow diagram of the search ordering method 200 of click model;
Fig. 3 shows click model schematic diagram according to an embodiment of the invention;
Fig. 4 shows the exemplary conceptual diagram of sort method according to an embodiment of the invention; And
Fig. 5 shows according to an embodiment of the invention based on the schematic diagram of the searching order device 500 of click model.
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.
Fig. 1 shows the exemplary environments 100 that search engine can run wherein.Environment 100 comprises by network 130, one or more clients 110 that such as the Internet, wide area network (WAN) or LAN (Local Area Network) (LAN) are connected to each other and one or more server 120 (normally " main frame ").Network 130 provides the access to the such as service of WWW (" web ") 131.
Web131 allows client 110 access to be included in the document based on text or content of multimedia in the webpage 121 (such as webpage or other documents) of such as being safeguarded by server 120 and serving.Usually, this is completed by the web browser application program 114 performed in client 110.The position of each page 121 can by being such as input in web browser application program 114 with accessed web page 121.Many webpages can be included in the hyperlink 123 of other webpages 121.Hyperlink 123 also can be the form of URL.Although the document that there has been described about the page realizes, be to be understood that environment 100 can comprise that have can by the content that characterizes and internuncial any link data object.
One skilled in the art will appreciate that, in general, search engine 140 corresponds to the online service of trustship in one or more computing machine and/or computing system, and wherein above-mentioned one or more computing machine or computing system are located and/or distributed in whole network 130.This search engine 140 receives and the search inquiry of customer in response end 110 submission.Especially, in response to inquiry, this search engine 140 obtain with receive search inquiry (being defined by the item of search inquiry) about and/or be correlated with search result information, i.e. result set 112.This result set 112 comprises Search Results, namely to from various different network site available about and/or relevant content quote (typically, form with hyperlink), wherein above-mentioned network site comprises the content hosting website of such as location in whole network 130.
As the skilled person will recognize, content hosting website trustship or to store by network 130 be available and/or addressable content for client 110 user.By using the process of the network sweep of climbing for content of grabbing, on multiple content hosting websites that search engine 140 will know location in whole network 130, the content of trustship at least partially.Once located content, this search engine 140 in content repository, will be equivalent to information bank 142, stores the information about trustship content.In response to inquiry, this search engine 140 extracts from information bank 142, returns the result set 112 of the item (such as keyword) meeting inquiry.
Because search engine 140 stores the page up to a million, especially when inquiry is loosely appointment, result set 112 can comprise many qualified pages.These pages can be relevant with the actual information demand of user or irrelevant.Therefore, the order of the result set 112 presented to client 110 affects the experience of user about search engine 140.
In one implementation, sequencer procedure can realize as a part for the ranking engine 144 in search engine 140, such as, and the searching order device 500 in this programme.In some implementations, sequencer procedure based on click logs, to improve the sequence of the page in result set 112, can more accurately can identify the page 113 relevant to specific topics like this.Finally, by information display device 146, the page in result set 112 is presented to user by the sequence improved.
Have been found that user more may click the higher page of rank, and no matter whether this page is in fact relevant to inquiry.This is called as position deviation.A kind of click model attempting to solve position deviation is position click model.This model hypothesis is only browsed result and is obtained a result and just click result when user is actual to when searching for relevant conclusion.That is, when user browses result and thinks that it is relevant, user only this result of perception is relevant, instead of really knows.Only when user's actual click result and browsing pages or document self time, whether user can understand result actual relevant.
The another kind of model distinguished between the reality and perceived relevance of result is series connection click model.For one query, series connection is clicked model hypothesis user and is browsed result successively in order, when user attract by certain result, user just clicks this result and inquires about, and a probability can be had to terminate inquiry no longer browse.
Although above-mentioned click model solves position deviation problem, user clicks behavior and can not explain by click information amount completely.Therefore, this programme proposes a kind of search ordering method based on click model, and the quantity of information reasonably user being clicked behavior is dissolved in the framework of searching order.
Fig. 2 shows according to an embodiment of the invention based on the process flow diagram of the search ordering method 200 of click model.The method starts from step S210, when a user query is received, obtain the first result sequence about user's inquiry, and each result items all has the first score value in this first result sequence, and all result items are according to the order sequence from high to low of the first score value.According to the embodiment of the present invention, the first score value calculates according at least one predetermined characteristic, and it is one or more that predetermined characteristic comprises in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity.That is, the first score value is the tolerance of the correlativity of result items and user search request in the first result sequence.
In addition, should understand, the first score value also can be that the feature calculation weighing search result relevance based on other obtains, and the present invention is not restricted predetermined characteristic.
Subsequently in step S220, utilize the second score value of each result items in the first result sequence in click model determining step S210, the second score value here, namely click feedback characteristic value.Like this, for each result items in the first result sequence, all there is first score value and second score value.According to the embodiment of the present invention, click model selects series connection click model.
The click result that click model uses, generally all from former pages of recent Search Results, before these results can come in sequence for the previous period, illustrates that they are most correlated results of thinking of sort algorithm at that time.Fig. 3 shows the schematic diagram of series connection click model, briefly, for one query q and result sequence R (r1, r2, rn), it is that to browse institute according to the order of r1 → r2 →...→ rn resultful that model hypothesis user is clicked in series connection, and a probability γ can be had to terminate inquiry no longer browse.In the process browsed, if user is attracted by result ri, user will click ri, and the probability of this behavior is ai.If user is satisfied to putting the ri result opened, user would not browse again, thus terminates current inquiry, and the probability of this behavior is si.So, ai and si of each result just constitute sequence A (a1, a2 ... an) and S (s1, s2 ..., sn), they and user above-mentioned abandon, together with the probability γ of search, constituting a probabilistic mathematical models, being called series connection click model.
Mention above, series connection click model can by obtaining the ability of user's click information, and the result making correlativity high is arranged more forward, but be not suitable for the weight determining separately sequence in search.Therefore, in searching order scheme of the present invention, the feature of traditional measurement correlativity and click result be merged mutually, the concrete method calculating weight order is shown in step S230.
Subsequently in step S230, choosing the second score value in the first result sequence is not the result items of 0, based on the order of the result items selected by the second score value adjustment, obtains the second result sequence, presents Query Result to user.According to above, because click behavior has openness, have the impact that a lot of result is not subject to click behavior, in the method, the second score value assignment this type of not being had affected result items is 0.
Give an example, if current, the sort algorithm of step S210 is thought has a result r should arrange before them, some features of certain explanation r are being improved in the recent period, such as viewing amount adds, article new especially, model are replied number and increased etc., r is made to have higher correlativity, before needing to be come.Now, if according to the sort result of click model, after r is probably come the result of click model impact, so just obliterated the lifting of this correlativity, that is, sequence produces positive feedback, and the result clicked namely can be more and more forward.So, need the sorting position retaining r, and then the measurement protecting other features (such as viewing amount, issuing time, money order receipt to be signed and returned to the sender quantity etc.) to promote correlativity, can not cover by clicked model.So in the method, only choose the result items that click model can have influence on, namely the second score value is not the result items of 0, resequence, and those second score values is the result items of 0 to them, keep their invariant positions in the first result sequence.
Afterwards, for the sequence of selected result items, the codomain character of click model be used.For the codomain of click model, generally with its mid point for boundary, upwards for forward divides, downward to divide for negative sense.Forward divides and shows that result should from current location toward front-seat, and negative sense divides and shows that result should be arranged from current location backward.In series connection click model, the forward of a result items divides larger, then it should past front-seat must be more; But negative sense is divided and there is no contrary conclusion.Negative sense divides and only represents that result is not suitable for coming current location, needs to move backward, but after the degree of moving be difficult to determine.
So in the method, the codomain average first choosing series connection click model as threshold value, be greater than codomain average for forward portions, remaining is negative sense portions, and will ensure that the second score value is that result items that forward divides comes before negative sense divides result items; Then, to there is result items that forward divides according to the order sequence from high to low of the second score value, now if there is the situation that the second score value is identical, just sort according to the first score value; Finally, to have the result items maintenance order that negative sense divides constant to all, here " maintenance order is constant " refers to, negative sense is kept to divide the relative ranks of result items constant by the order of the first score value in step S210, so just ensure that the result items that negative sense divides can not can not get clicking because of being pressed into very below, thus difficultly again to be come up by row.
So far, just obtain the second result sequence after to the first result sequence rearrangement, in this time, by the first score value of the result items that position in the second result sequence changes, be adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items.That is, in the second result sequence, the first score value still to be kept to be present by order from high to low.So just maintain the second result sequence still according to the character that the first score value sorts from big to small, make all multioperations based on this character in search engine, such as this sequence and the result sequence of other parallel search engines are combined together minor sort again, still effective after use this method.
In order to definitely for the explanation of this method 200, Fig. 4 shows the exemplary conceptual diagram of sort method according to an embodiment of the invention.
Wherein the sequence of the leftmost side represents the first result sequence according to the first score value sequence, every in first result sequence is all one and comprises < result items sequence number, first score value, the tlv triple of the second score value >, and set the codomain of the first score value and the second score value all as [0,100], so codomain average is 50.The result items that the second score value is not 0 is selected from the first result sequence, obtain a < result items sequence number, two tuples of the second score value >, as <2, 11>, <3, 99> ... then by mentioned above, only change the order that the second score value is greater than the result items of codomain average 50, and the result items that the second score value is not more than 50 keeps former order constant, and ensure result items that the second score value is greater than 50 all come be not more than 50 result items before.Such as, the relative ranks of item 2 and item 6 does not just change.Finally, according to sequencing, successively the result items sequence number of the above results item and the second score value being put back into the second score value in the first result sequence is not in the position corresponding to result items of 0, and does not change the position that the second score value is the result items of 0, obtains the second result sequence on right side.Note, the first result sequence is identical with the first score value of the result items of the second result sequence correspondence position, and that is, their the first score value all maintains sequence from high to low.
Utilize this method 200, on the basis of original collating sequence (i.e. the first collating sequence), in conjunction with the model feature of series connection click model, reasonably power is adjusted to be dissolved in the framework of searching order click, change position and the weights of result items, make the final DISPLAY ORDER of Query Result, not only reflect the positive effect of click model to correlativity, also avoid positive feedback shortcoming.
For checking the effect of this method 200, this method is applied, with DBN (dynamicbayesiannetwork, dynamic bayesian network in forum's search engine of the family of automobile, a kind of tandem type click model) be click model, calculate the returning rate of after applying this method several days.Returning rate refers to the user using next product, uses its ratio within a period of time again, is generally used for weighing user to the viscosity of this product or favorable rating.In simple terms, search engine sequence is better, and nature can obtain higher returning rate; If click tune power to cause positive feedback, what user was seen at former page is the result that those were clicked all the time, and returning rate can transfer decline, therefore higher returning rate shows that positive feedback problem is less.The computing formula of returning rate is as follows:
If user employed product at the 0th day, then
The returning rate of n-th day=
| ((user of the 1st day) ∪ (user of the 2nd day) ... ∪ (user of n-th day)) ∩ (user of the 0th day) | ÷ | the user of the 0th day |
Wherein, ∩ asks friendship, and ∪ asks also, | x| represents the quantity of x.
Use small amount data compares, and the product returning rate of first four days use/unuse this method is compared as follows continuously:
Use this method | Do not use this method | Lifting ratio |
8.4% | 6.6% | 27.3% |
12.1% | 9.6% | 26.0% |
19.1% | 14.7% | 29.9% |
23.0% | 17.7% | 29.9% |
Visible, this method is obvious to the lifting of returning rate, illustrates that the present invention under the prerequisite not causing larger problem, can effectively promote the correlativity of Search Results.
Fig. 5 shows according to an embodiment of the invention based on the schematic diagram of the searching order device 500 of click model.This device comprises: acquiring unit 510, computing unit 520 and sequencing unit 530.Wherein computing unit 520 couples mutually with acquiring unit 510 and sequencing unit 530 respectively.
Acquiring unit 510 is suitable for obtaining the first result sequence about user's inquiry, and wherein in this first result sequence, each result items has the first score value.This first score value is calculated by computing unit 520, and computing unit 520 is suitable for calculating the first score value according at least one predetermined characteristic, and according to the embodiment of the present invention, it is one or more that predetermined characteristic comprises in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity.Further, sequencing unit 530 is suitable for all result items in the first result sequence according to the order sequence from high to low of the first score value.
Then, computing unit 520 is also suitable for utilizing click model to determine the second score value of each result items in the first result sequence.According to one embodiment of present invention, click model is series connection click model.
It is not the result items of 0 that sequencing unit 530 is also suitable for choosing the second score value in described first result sequence, based on the order of the result items selected by the second score value adjustment, to obtain the second result sequence.Particularly, it is not in the result items of 0 that sequencing unit 530 also comprises the second score value be suitable for selected by judgement, and whether the second score value is greater than the judgment sub-unit 532 of threshold value.According to the embodiment of the present invention, this threshold value is the codomain average of series connection click model.Then sequencing unit 530 is also suitable for the result items the second score value being greater than to threshold value, according to the order sequence from high in the end of the second score value; And the second score value is not more than to the result items of threshold value, maintenance order is constant, and come after the second score value is greater than all result items of threshold value, here " maintenance order is constant " refers to, the second score value is kept to be not more than the relative ranks of the result items of threshold value by the order of the first score value constant, this is done to ensure that the result items that negative sense divides can not can not get clicking because of being pressed into very below, thus difficultly again to be come up by row.Finally, sequencing unit 530 is also suitable for the first score value of the result items changed position in the second result sequence, be adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items, do like this and just ensure that in the second result sequence, the first score value is also by order sequence from big to small, make all multioperations based on this ordering property in search engine, such as this sequence and the result sequence of other parallel search engines are combined together minor sort again, still effective after use this method.
About concrete steps and the embodiment of sequence, open in detail in based on the description of Fig. 4, repeat no more herein.
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 than the feature more multiple features clearly recorded in each claim.Or rather, as claims below 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 to be understood that the module of the equipment in example disclosed herein or unit or assembly can be arranged in equipment as depicted in this embodiment, or alternatively can be positioned in one or more equipment different from the equipment in this example.Module in aforementioned exemplary can be combined as a module or can be divided into multiple submodule in addition.
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.
A6, method as described in A5, wherein also comprise: by the first score value of the result items that position in the second result sequence changes, be adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items.
B11, device as described in B10, wherein threshold value is the codomain average of series connection click model.B12, device as described in B11, wherein, sequencing unit is also suitable for the first score value of the result items changed position in the second result sequence, is adjusted to the first score value of result items in the first result sequence corresponding to the existing position of this result items.
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 the following claims, the one of any of embodiment required for protection can use with arbitrary array mode.
In addition, some in described embodiment are described as at this can by the processor of computer system or the method implemented by other device performing described function or the combination of method element.Therefore, there is the device of processor formation for implementing the method or method element of the necessary instruction for implementing described method or method element.In addition, the element described herein of device embodiment is the example as lower device: this device is for implementing the function performed by the element of the object in order to implement this invention.
As used in this, unless specifically stated so, use ordinal number " first ", " second ", " the 3rd " etc. to describe plain objects and only represent the different instances relating to similar object, and be not intended to imply the object be described like this must have the time upper, spatially, sequence aspect or in any other manner to definite sequence.
Although the embodiment according to limited quantity describes the present invention, benefit from description above, those skilled in the art understand, in the scope of the present invention described thus, it is contemplated that other embodiment.In addition, it should be noted that the language used in this instructions is mainly in order to object that is readable and instruction is selected, instead of select to explain or limiting theme of the present invention.Therefore, when not departing from the scope and spirit of appended claims, many modifications and changes are all apparent for those skilled in the art.For scope of the present invention, be illustrative to disclosing of doing of the present invention, and nonrestrictive, and scope of the present invention is defined by the appended claims.
Claims (10)
1., based on a search ordering method for click model, described method comprises step:
Obtain the first result sequence about inquiry, in described first result sequence, each result items has the first score value, and all result items are according to the order sequence from high to low of the first score value, and described first score value calculates according at least one predetermined characteristic;
Click model is utilized to determine the second score value of each result items in described first result sequence; And
Choosing the second score value in described first result sequence is not the result items of 0, based on the order of the result items selected by the second score value adjustment, obtains the second result sequence.
2. the method for claim 1, wherein described predetermined characteristic to comprise in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity one or more.
3. method as claimed in claim 1 or 2, wherein, described click model is series connection click model.
4. method as claimed in claim 3, the step of the wherein said order based on the result items selected by the second score value adjustment comprises:
Second score value is greater than to the result items of threshold value, according to the order sequence from high in the end of the second score value;
Second score value is not more than to the result items of threshold value, maintenance order is constant, and comes after the second score value is greater than all result items of threshold value.
5. method as claimed in claim 4, wherein said threshold value is the codomain average of series connection click model.
6., based on a searching order device for click model, described device comprises:
Acquiring unit, be suitable for obtaining the first result sequence about inquiry, in wherein said first result sequence, each result items has the first score value;
Computing unit, is suitable for calculating the first score value according at least one predetermined characteristic, being also suitable for utilizing click model to determine the second score value of each result items in described first result sequence; And
Sequencing unit, being suitable for all result items in described first result sequence to sort according to the first score value order from high to low, be also suitable for choosing the second score value in described first result sequence is not the result items of 0, based on the order of the result items selected by the second score value adjustment, to obtain the second result sequence.
7. device as claimed in claim 6, wherein, it is one or more that described predetermined characteristic comprises in viewing amount, issuing time and money order receipt to be signed and returned to the sender quantity.
8. device as claimed in claims 6 or 7, wherein said click model is series connection click model.
9. device as claimed in claim 8, wherein,
Described sequencing unit also comprises judgment sub-unit, and the second score value be suitable for selected by judging is not that in the result items of 0, whether the second score value is greater than threshold value;
Described sequencing unit is also suitable for the result items the second score value being greater than to threshold value, according to the order sequence from high in the end of the second score value; And the second score value being not more than to the result items of threshold value, maintenance order is constant, and comes after the second score value is greater than all result items of threshold value.
10. an information search engine system, comprising:
Information bank, is suitable for storing information to be put;
The searching order device based on click model according to any one of claim 6-9, is suitable for sorting to inquiring about the result sequence obtained; And
Information display device, is suitable for showing Query Result in order.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510697625.2A CN105302898B (en) | 2015-10-23 | 2015-10-23 | A kind of search ordering method and device based on click model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510697625.2A CN105302898B (en) | 2015-10-23 | 2015-10-23 | A kind of search ordering method and device based on click model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105302898A true CN105302898A (en) | 2016-02-03 |
CN105302898B CN105302898B (en) | 2019-02-19 |
Family
ID=55200168
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510697625.2A Active CN105302898B (en) | 2015-10-23 | 2015-10-23 | A kind of search ordering method and device based on click model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105302898B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107273112A (en) * | 2017-05-04 | 2017-10-20 | 武汉斗鱼网络科技有限公司 | A kind of present list method for information display and device |
CN110598089A (en) * | 2018-05-25 | 2019-12-20 | 重庆好德译信息技术有限公司 | Voice assistant method and system based on environmental information processing |
CN110825939A (en) * | 2019-09-19 | 2020-02-21 | 五八有限公司 | Method and device for generating and sorting scores of posts, electronic equipment and storage medium |
CN111905378A (en) * | 2020-08-19 | 2020-11-10 | 上海莉莉丝网络科技有限公司 | Data updating system, data updating method and server |
CN113254810A (en) * | 2021-06-17 | 2021-08-13 | 浙江口碑网络技术有限公司 | Search result output method and device, computer equipment and readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090019036A1 (en) * | 2007-07-10 | 2009-01-15 | Asim Roy | Systems and Related Methods of User-Guided Searching |
CN101930438A (en) * | 2009-06-19 | 2010-12-29 | 阿里巴巴集团控股有限公司 | Search result generating method and information search system |
CN102004782A (en) * | 2010-11-25 | 2011-04-06 | 北京搜狗科技发展有限公司 | Search result sequencing method and search result sequencer |
CN103593353A (en) * | 2012-08-15 | 2014-02-19 | 阿里巴巴集团控股有限公司 | Information search method and display information sorting weight value determination method and device |
CN103970796A (en) * | 2013-02-04 | 2014-08-06 | 深圳市世纪光速信息技术有限公司 | Inquiry preference ordering method and device |
-
2015
- 2015-10-23 CN CN201510697625.2A patent/CN105302898B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090019036A1 (en) * | 2007-07-10 | 2009-01-15 | Asim Roy | Systems and Related Methods of User-Guided Searching |
CN101930438A (en) * | 2009-06-19 | 2010-12-29 | 阿里巴巴集团控股有限公司 | Search result generating method and information search system |
CN102004782A (en) * | 2010-11-25 | 2011-04-06 | 北京搜狗科技发展有限公司 | Search result sequencing method and search result sequencer |
CN103593353A (en) * | 2012-08-15 | 2014-02-19 | 阿里巴巴集团控股有限公司 | Information search method and display information sorting weight value determination method and device |
CN103970796A (en) * | 2013-02-04 | 2014-08-06 | 深圳市世纪光速信息技术有限公司 | Inquiry preference ordering method and device |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107273112A (en) * | 2017-05-04 | 2017-10-20 | 武汉斗鱼网络科技有限公司 | A kind of present list method for information display and device |
CN107273112B (en) * | 2017-05-04 | 2021-02-02 | 武汉斗鱼网络科技有限公司 | Method and device for displaying gift list information |
CN110598089A (en) * | 2018-05-25 | 2019-12-20 | 重庆好德译信息技术有限公司 | Voice assistant method and system based on environmental information processing |
CN110825939A (en) * | 2019-09-19 | 2020-02-21 | 五八有限公司 | Method and device for generating and sorting scores of posts, electronic equipment and storage medium |
CN110825939B (en) * | 2019-09-19 | 2023-10-13 | 五八有限公司 | Post score generation and ordering method and device, electronic equipment and storage medium |
CN111905378A (en) * | 2020-08-19 | 2020-11-10 | 上海莉莉丝网络科技有限公司 | Data updating system, data updating method and server |
CN111905378B (en) * | 2020-08-19 | 2024-04-02 | 上海莉莉丝网络科技有限公司 | Data updating system, data updating method and server |
CN113254810A (en) * | 2021-06-17 | 2021-08-13 | 浙江口碑网络技术有限公司 | Search result output method and device, computer equipment and readable storage medium |
WO2022262849A1 (en) * | 2021-06-17 | 2022-12-22 | 浙江口碑网络技术有限公司 | Search result output method and apparatus, computer device and readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN105302898B (en) | 2019-02-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
AU2011202345B2 (en) | Methods and systems for improving a search ranking using related queries | |
US8645390B1 (en) | Reordering search query results in accordance with search context specific predicted performance functions | |
US10534781B2 (en) | Website traffic optimization | |
US8515937B1 (en) | Automated identification and assessment of keywords capable of driving traffic to particular sites | |
JP5328896B2 (en) | Query refinement and proposals using social networks | |
CN105302898A (en) | Click model-based searching and ranking method and device | |
US8478746B2 (en) | Operationalizing search engine optimization | |
CN101266620B (en) | Method and equipment for providing target information for user | |
US20130179428A1 (en) | Method and system for ranking results and providing lists of experts from social networks | |
US8195654B1 (en) | Prediction of human ratings or rankings of information retrieval quality | |
US20070168342A1 (en) | Recommended alerts | |
US9691083B2 (en) | Opportunity identification and forecasting for search engine optimization | |
US9020922B2 (en) | Search engine optimization at scale | |
AU2006332534A1 (en) | Predicting ad quality | |
US8423554B2 (en) | Content category scoring for nodes in a linked database | |
CN103309894B (en) | Based on search implementation method and the system of user property | |
WO2013126598A1 (en) | Determining search result rankings based on trust level values associated with sellers | |
US20140108376A1 (en) | Enhanced detection of like resources | |
US20210191995A1 (en) | Generating and implementing keyword clusters | |
US20090248655A1 (en) | Method and Apparatus for Providing Sponsored Search Ads for an Esoteric Web Search Query | |
WO2016018467A1 (en) | Behavior influenced search ranking | |
Karthikeyan et al. | Page rank based design and implementation of search engine optimization | |
Jagithyala | Recommending recipes based on ingredients and user reviews | |
Najafi et al. | A New Hybrid Method for Web Pages Ranking in Search Engines | |
JP6100677B2 (en) | Information recommendation device, information recommendation method, and information recommendation program |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20180914 Address after: 100089 Beijing Haidian District Haidian District Dan Street 3 B block 11, 1110, 1111 rooms. Applicant after: Che Zhi interconnect (Beijing) Technology Co., Ltd. Address before: 300300 Tianjin Binhai New Area Airport International Logistics Area Second Street 1 Enterprise Service Center 311 room. Applicant before: TIANJIN CHESHIJIA TECHNOLOGY CO., LTD. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |