CN105069077A - Search method and device - Google Patents
Search method and device Download PDFInfo
- Publication number
- CN105069077A CN105069077A CN201510463375.6A CN201510463375A CN105069077A CN 105069077 A CN105069077 A CN 105069077A CN 201510463375 A CN201510463375 A CN 201510463375A CN 105069077 A CN105069077 A CN 105069077A
- Authority
- CN
- China
- Prior art keywords
- behavior
- user
- intention
- result
- historical search
- 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.)
- Pending
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 embodiment of the invention discloses a search method and device. The method comprises the following steps: identifying an intention category to which a target search statement input by a current user belongs; according to the intention category obtained by identification and a user behavior model established in advance, predicting the behaviors of the current user on candidate search results, wherein the user behavior model is determined according to the intention category to which a historical search statement of a historical user belongs; and on the basis of a prediction result, determining a current output search result in the candidate search results. The technical scheme provided by the embodiment of the invention can use the reasonable and accurate user behavior model, provides the high-quality search result for the user and improves a search satisfaction degree of the user.
Description
Technical field
The embodiment of the present invention relates to Internet technical field, particularly relates to searching method and device.
Background technology
Along with the continuous upgrading of network infrastructure, for solving problem of information overload, give birth to for providing the search engine of search service to apply for user.At present; under line understood usually by search engine, study in advance obtains personal behavior model; and after receiving the search statement (Query) that user inputs in real time on line; utilize personal behavior model to carry out user behavior understanding based on this search statement, and then determine Search Results in conjunction with the understanding result of user behavior.Wherein, personal behavior model for predicting the behavior of user to webpage, such as predict user whether webpage clicking, browse the duration of webpage or the satisfaction etc. to webpage.
But, prior art is only based on the attributive character (such as location, age, sex, occupation) of user, text feature (the historical search statement such as inputted), search behavior (such as whether click, browse duration) etc. to historical search result, sets up personal behavior model.Although these features cover extensive, abundant in content, inherently do not understand modeling to user behavior, make to set up the personal behavior model obtained reasonable not, well can not understand the search need of user, Consumer's Experience is poor.
Summary of the invention
The embodiment of the present invention provides a kind of searching method and device, can utilize comparatively rationally personal behavior model accurately, for user provides the Search Results of high-quality more, promotes the satisfaction of user to search.
On the one hand, embodiments provide a kind of searching method, the method comprises:
Identify the intention classification belonging to target search statement of active user's input;
According to identifying the intention classification obtained and the personal behavior model be pre-created, active user is to the behavior of candidate search result in prediction, and the intention classification of wherein said personal behavior model belonging to the historical search statement of historic user is determined;
Based on predicting the outcome, from described candidate search result, determine this Search Results exported.
On the other hand, the embodiment of the present invention additionally provides a kind of searcher, and this device comprises:
Intention classification recognition unit, for identifying the intention classification belonging to the target search statement that active user inputs;
User's behavior prediction unit, for the intention classification obtained according to the identification of described intention classification recognition unit and the personal behavior model be pre-created, active user is to the behavior of candidate search result in prediction, and the intention classification of wherein said personal behavior model belonging to the historical search statement of historic user is determined;
Search result output unit, for predicting the outcome of obtaining based on described user's behavior prediction unit, determines this Search Results exported from described candidate search result.
The technical scheme that the embodiment of the present invention provides, user behavior modeling is carried out based on the intention belonging to the historical search statement that historic user inputs, deep enoughly can understand user behavior, make personal behavior model more reasonable accurately, and then to utilize on this model prediction line user to the behavior of candidate search result, the Search Results of high-quality more can be provided for user, promote the satisfaction of user to search.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of a kind of searching method that the embodiment of the present invention one provides;
Fig. 2 is the schematic flow sheet of a kind of searching method that the embodiment of the present invention four provides;
Fig. 3 is the structural representation of a kind of searcher that the embodiment of the present invention five provides.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further detail.Be understandable that, specific embodiment described herein is only for explaining the present invention, but not limitation of the invention.It also should be noted that, for convenience of description, illustrate only part related to the present invention in accompanying drawing but not entire infrastructure.
Before in further detail exemplary embodiment being discussed, it should be mentioned that some exemplary embodiments are described as the process or method described as process flow diagram.Although operations (or step) is described as the process of order by process flow diagram, many operations wherein can be implemented concurrently, concomitantly or simultaneously.In addition, the order of operations can be rearranged.Described process can be terminated when its operations are completed, but can also have the additional step do not comprised in the accompanying drawings.Described process can correspond to method, function, code, subroutine, subroutine etc.
Day by day perfect along with network infrastructure, Internet resources are more and more abundanter, and all multi-users can have search need miscellaneous.For meeting this search need, user normally intentionally inputs corresponding search statement by search box and searches for, and the Search Results returned search engine afterwards is such as clicked, browse, the behavior of collection etc. and so on.These user behaviors have close association with intention motivation.
At present, search engine is after receiving the search statement that any user inputs in real time, can based on this search statement, the personal behavior model established in advance under utilizing line, user is to the behavior outcome of each candidate search result in prediction, and then predicts the outcome from multiple candidate search result, select the Search Results of this secondary output based on this.Wherein, described behavior outcome is normally for the metric parameter values of characterizing consumer to candidate result satisfaction.
But in prior art, scheme is set up to personal behavior model under line, can not consider this factor of intention motivation, user behavior understanding is carried out by intention analysis, but simply regardless of intention by general behavioural characteristic corresponding for the historical search result of historic user under historical search statement and general behavior outcome, as train example carry out modeling.So, prior art does not inherently understand modeling to user behavior, thus cause setting up the personal behavior model that obtains can not Accurate Prediction user to the behavior outcome of candidate search result because the behavior outcome of user and search intention have inseparable relation.
So-called general behavioural characteristic refers to: the behavioural characteristic not comprising the historic user of intent features.Exemplary, general behavioural characteristic can comprise following at least one feature usually: user property feature (such as comprising user location, age, sex, occupation etc.), user's input feature vector (the historical search statement such as inputted), the text feature (theme, the keyword of webpage corresponding to such as historical search result) etc. of historical search result.
So-called general behavior outcome refers to: have nothing to do with intent features, adopt unified for characterizing the value that the metric parameter of historic user to historical search result satisfaction obtains.Also namely, no matter historic user is the search intent what is intended to, and what adopt is all identical for characterizing the metric parameter of historic user to historical search result satisfaction.Such as, for any search intention, metric parameter is the residence time of historic user to historical search result.But under the search intention of historic user is info class intention, the residence time of historic user to be satisfied with historical search result is often longer; And under the search intention of historic user is resource downloading class intention, historic user is but very short to the residence time of be satisfied with historical search result, only has the click of download link for several times.Obviously, the feature of these user behaviors has close association with search intention, the performance of the user behavior that different search intention triggers when satisfied to Search Results is inconsistent, thus unified metric parameter is adopted to carry out characterizing consumer careful not to candidate result satisfaction, just cause the comprehension of the personal behavior model of foundation to user behavior poor accordingly, accuracy is not high yet.
In view of this, the embodiment of the present invention introduces the thought that intention is analyzed, and sets up personal behavior model based on intent features, and then to utilize on this model prediction line user to the behavior of candidate search result, think that user provides the Search Results of high-quality more, promote the satisfaction of user to search.
Embodiment one
Fig. 1 is the schematic flow sheet of a kind of searching method that the embodiment of the present invention one provides.The method of the present embodiment can be performed by searcher, and this device, by software simulating, is integrated in as terminal device miscellaneous provides in the search engine of search service.See Fig. 1, the operation included by the searching method that the present embodiment provides is specific as follows.
The intention classification belonging to target search statement of S110, identification active user input.
The search statement comprised in the searching request sent by the search client in terminal or browser that target search statement receives for search engine.This search statement can be the phonetic search information that active user inputs in phonetic entry mode, or with the text search content that Text Input mode inputs.
After getting target search statement, based on the intention classification recognizer of setting, this search statement is identified, to determine which class intention classification this search statement specifically belongs to.It should be noted that, the present embodiment is not specifically limited the kind be intended to and number.Preferably, be intended to classification and comprise following at least one: addressing class intention (also namely searching the intention of URL(uniform resource locator)), info class intention (the inquiry information of problem or the intention of solution) and commodity class intention (searching the intention of merchandise news).Exemplary, multiple intention classification can be preset, and the mapping relations set up between intention classification and keyword are stored.Wherein, often kind of intention classification can be mapped with multiple keyword simultaneously.Such as, the keyword being intended to have mapping relations with addressing class has: " website ", " network address " etc.; The keyword being intended to have mapping relations with info class has: " why ", " how ", " how ", " whether " etc.; The keyword being intended to have mapping relations with commodity class has: " Jingdone district ", " Taobao ", " sky cat ", " masses comment on " etc.
Accordingly, identify the intention classification belonging to target search statement of active user's input, can specifically comprise: the target search statement of active user's input is resolved, to extract wherein comprised keyword; From the mapping relations between the intention classification be pre-created and keyword, inquiry extracts the intention classification that the keyword obtained has mapping relations, using this intention classification as the intention classification belonging to target search statement with this.
S120, according to identifying the intention classification that obtains and the personal behavior model be pre-created, active user is to the behavior of candidate search result in prediction.
In the present embodiment, can personal behavior model be pre-created and store.The intention classification of this personal behavior model belonging to the historical search statement of historic user is determined.Exemplary, the constructive process of personal behavior model, specifically comprises: the historical behavior data obtaining historic user; Identify the intention classification belonging to historical search statement in historical behavior data; Intention classification belonging to historical search statement and historical behavior data, set up personal behavior model.Wherein, to the recognition technology being intended to classification belonging to historical search statement, identical with the above-mentioned recognition technology to being intended to classification belonging to target search statement.
In this example, need the corpus being pre-created the historical behavior data including many historic user, the method based on machine learning utilizes this training to obtain personal behavior model.The historical behavior data of each historic user can comprise: the attributive character data of this user; The historical search statement that this user inputted; The text feature data of each historical search result that search engine returns according to historical search statement; This user to the behavior record of each historical search result, the behavior record can be record to comprising following at least one content: whether click historical search result, stay time in webpage corresponding to historical search result and whether collect this webpage.
As a kind of embodiment of the present embodiment, can respectively for various intention classification under line, be pre-created a personal behavior model corresponding with this intention classification.At the target search statement getting active user's input on line, and after identifying the intention classification belonging to it, this that be pre-created is utilized to identify this personal behavior model corresponding to intention classification obtained, predict that active user is to the behavior of each candidate search result, to obtain behavior outcome respectively.
As the another kind of embodiment of the present embodiment, a general personal behavior model can be pre-created under line, should increase one dimension intent features when setting up this personal behavior model, wherein this is characterized as the feature for characterizing the intention classification belonging to historical search statement corresponding to training example.At the target search statement acquiring active user's input on line, and after identifying the intention classification obtained belonging to it, generate intent features, the personal behavior model be pre-created then is utilized according to this intent features, predict that active user is to the behavior of each candidate search result, to obtain behavior outcome respectively.
S130, based on predicting the outcome, from candidate search result, determine this Search Results exported.
Exemplary, can predict the outcome according to corresponding, each candidate search result is sorted; The candidate search result of front N (for being more than or equal to the natural number of 1) name will be positioned at, as the Search Results that this exports.
The technical scheme that the present embodiment provides, user behavior modeling is carried out based on the intention belonging to the historical search statement that historic user inputs, deep enoughly can understand user behavior, make personal behavior model more reasonable accurately, and then to utilize on this model prediction line user to the behavior of candidate search result, the Search Results of high-quality more can be provided for user, promote the satisfaction of user to search.
Embodiment two
The present embodiment, on the basis of above-described embodiment one, is optimized the foundation of personal behavior model.In the present embodiment, the corresponding personal behavior model of each intention classification.Concrete, the intention classification belonging to historical search statement and historical behavior data, set up personal behavior model, comprising:
From historical behavior data, be extracted in user behavior feature corresponding to historical search result under historical search statement and user behavior result, wherein the intention classification of user behavior result belonging to historical search statement is determined;
Using user behavior characteristic sum user behavior result as training example, set up the personal behavior model corresponding with the intention classification belonging to historical search statement based on machine learning algorithm.
Wherein, the user behavior feature extracted, can be the Partial Feature in the general behavioural characteristic in prior art or whole feature, such as, comprises: the attributive character of historic user, user's input feature vector corresponding to historical search statement of input, and the text feature etc. of historical search result.
The user behavior result extracted, is different from the general behavior outcome in prior art, is the behavior outcome corresponding with the intention classification belonging to historical search statement.Under difference intention classification, what adopt can be identical for characterizing the metric parameter of historic user to historical search result satisfaction, also can be different.Such as, under info class intention, metric parameter can be the residence time of historic user to historical search result, and this time is longer, shows that user is more satisfied; And under resource downloading class intention, whether metric parameter can be clicked historical search result for historic user.
Getting the target search statement of active user's input, and after identifying the intention classification obtained belonging to it, for each candidate search result, perform and operate as follows: determine the user behavior feature that the candidate search result of active user under target search statement is corresponding; The user behavior feature utilizing this to determine and be pre-created this identify this personal behavior model corresponding to intention classification obtained, active user is to the behavior of candidate search result, to obtain behavior outcome in prediction.
Embodiment three
The present embodiment, on the basis of above-described embodiment one, is optimized the foundation of personal behavior model.In the present embodiment, all corresponding same general personal behavior model of various intention classification.Being different from prior art, one dimension intent features should being increased when setting up this personal behavior model.Concrete, can using intention classification corresponding for historical search statement as the feature (abbreviation intent features) in a dimension of personal behavior model, and using corresponding with intent features, be used for characterizing the value of historic user to the metric parameter of historical search result satisfaction as user behavior result, to form training example; And then utilize training example to train based on the method for machine learning, set up general personal behavior model.
Intention classification belonging to historical search statement and historical behavior data, set up personal behavior model, comprising:
From historical search behavioral data, be extracted in user behavior feature corresponding to historical search result under historical search statement and user behavior result, wherein the intention classification of user behavior result belonging to described historical search statement is determined;
Using the intention category feature belonging to user behavior feature, historical search statement and user behavior result as training example, set up personal behavior model based on machine learning algorithm.
Wherein, the user behavior feature extracted, can be the Partial Feature in the general behavioural characteristic in prior art or whole feature, such as, comprises: the attributive character of historic user, user's input feature vector corresponding to historical search statement of input, and the text feature etc. of historical search result.
The user behavior result extracted, is different from the general behavior outcome in prior art, is the behavior outcome corresponding with the intention classification belonging to historical search statement.Under difference intention classification, what adopt can be identical for characterizing the metric parameter of historic user to historical search result satisfaction, also can be different.
Getting the target search statement of active user's input, and after identifying the intention classification obtained belonging to it, for each candidate search result, performing and operate as follows: generate intent features according to recognition result; Determine the user behavior feature that the candidate search result of active user under target search statement is corresponding; The user behavior feature utilizing this to determine, intent features and the general personal behavior model be pre-created, active user is to the behavior of candidate search result, to obtain behavior outcome in prediction.
Embodiment four
Fig. 2 is the schematic flow sheet of a kind of searching method that the embodiment of the present invention four provides.The present embodiment, based on above-described embodiment, provides a preferred embodiment.Be intended to this 3 class with addressing class intention, info class intention and commodity class be in the present embodiment intended to example and be introduced, but those of ordinary skill in the art should understand, in the middle of reality performs, intention classification can also be other dividing mode, and the present embodiment is explained explanation to technical scheme provided by the invention in an illustrative manner.See Fig. 2, the operation included by the searching method that the present embodiment provides is specific as follows.
Before this, the operation of at least one group of historical behavior data obtaining multiple historic user is performed.Wherein, each group historical behavior data is to having: the attributive character data of historic user; A historical search statement of input; The text feature data of each historical search result that search engine returns according to this historical search statement; This user to the behavior record of each historical search result, the behavior record can be record to comprising following at least one content: whether click historical search result, stay time in webpage corresponding to historical search result and whether collect this webpage.
Then, perform according to wherein comprised historical search statement, to the operation often organizing historical behavior data and carry out being intended to category division.Concrete, based on the mapping relations between the intention classification be pre-created and keyword, identify the intention classification belonging to each bar historical search statement; Be the historical behavior data under this identifies the intention classification obtained by this historical search statement place historical behavior Data Placement.
Subsequently, historical behavior data under each intention classification: be therefrom extracted in user behavior feature corresponding to each historical search result under historical search statement and user behavior result, using the user behavior characteristic sum user behavior result of extraction as training example, carry out the user behavior modeling of each intention classification based on machine learning algorithm.Wherein, the intention classification of user behavior result belonging to historical search statement is determined.
And then, obtain the personal behavior model that each intention classification is corresponding: the personal behavior model that the personal behavior model of addressing class intention correspondence, info class intention are corresponding and commodity class are intended to corresponding personal behavior model.
Above-mentioned all operations can online under perform in advance.
On line, receive the target search statement (being also new Query) of active user's input before this, identify the intention classification belonging to new Query; Then according to the personal behavior model corresponding with recognition result, active user is to the behavior of each candidate search result in prediction, and then from each candidate search result, determines this Search Results exported based on predicting the outcome.Concrete forecasting process see the associated description in above-described embodiment two, can not repeat them here.
The intention that the present embodiment can be correlated with for Query carries out more careful modeling, and then promote user behavior understanding, therefore the effect that user behavior is understood can be promoted: the Search Results providing high-quality more for user, finally promotes satisfaction of users.
Embodiment five
Fig. 3 is the structural representation of a kind of searcher that the embodiment of the present invention five provides.See Fig. 3, the concrete structure of this device is as follows:
Intention classification recognition unit 310, for identifying the intention classification belonging to the target search statement that active user inputs;
User's behavior prediction unit 320, for identifying the intention classification obtained and the personal behavior model be pre-created according to intention classification recognition unit 310, active user is to the behavior of candidate search result in prediction, and wherein the intention classification of personal behavior model belonging to the historical search statement of historic user is determined;
Search result output unit 330, for predicting the outcome of obtaining based on user's behavior prediction unit 320, determines this Search Results exported from candidate search result.
Exemplary, the searcher that the present embodiment provides also comprises personal behavior model and sets up unit 300, and wherein personal behavior model is set up unit 300 and comprised:
Data acquisition subelement 301, for obtaining the historical behavior data of historic user;
Intention assessment subelement 302, for identifying the intention classification belonging to the historical search statement in historical behavior data;
Subelement 303 set up by model, for the intention classification belonging to historical search statement and historical behavior data, sets up personal behavior model.
As a kind of embodiment of the present embodiment, subelement 303 set up by described model, specifically for:
From historical behavior data, be extracted in user behavior feature corresponding to historical search result under historical search statement and user behavior result, the intention classification of wherein said user behavior result belonging to described historical search statement is determined;
Using user behavior characteristic sum user behavior result as training example, set up the personal behavior model corresponding with the intention classification belonging to historical search statement based on machine learning algorithm.
As the another kind of embodiment of the present embodiment, subelement 303 set up by model, specifically for:
From historical search behavioral data, be extracted in user behavior feature corresponding to historical search result under historical search statement and user behavior result, wherein the intention classification of user behavior result belonging to described historical search statement is determined;
Using the intention category feature belonging to user behavior feature, historical search statement and described user behavior result as training example, set up personal behavior model based on machine learning algorithm.
On the basis of the technique scheme provided at the present embodiment, intention classification comprises following at least one: addressing class intention, info class intention and commodity class intention.
This product of above-mentioned searcher can perform the method that any embodiment of the present invention provides, possess the corresponding functional module of manner of execution and beneficial effect, the not ins and outs of detailed description in the present embodiment, can see the embodiment for interpret-execution method in the embodiment of the present invention.
Note, above are only preferred embodiment of the present invention and institute's application technology principle.Skilled person in the art will appreciate that and the invention is not restricted to specific embodiment described here, various obvious change can be carried out for a person skilled in the art, readjust and substitute and can not protection scope of the present invention be departed from.Therefore, although be described in further detail invention has been by above embodiment, the present invention is not limited only to above embodiment, when not departing from the present invention's design, can also comprise other Equivalent embodiments more, and scope of the present invention is determined by appended right.
Claims (10)
1. a searching method, is characterized in that, comprising:
Identify the intention classification belonging to target search statement of active user's input;
According to identifying the intention classification obtained and the personal behavior model be pre-created, active user is to the behavior of candidate search result in prediction, and the intention classification of wherein said personal behavior model belonging to the historical search statement of historic user is determined;
Based on predicting the outcome, from described candidate search result, determine this Search Results exported.
2. method according to claim 1, is characterized in that, also comprises:
Obtain the historical behavior data of historic user;
Identify the intention classification belonging to historical search statement in described historical behavior data;
Intention classification belonging to described historical search statement and described historical behavior data, set up personal behavior model.
3. method according to claim 2, is characterized in that, the intention classification belonging to described historical search statement and described historical behavior data, sets up personal behavior model, comprising:
From described historical behavior data, be extracted in user behavior feature corresponding to historical search result under described historical search statement and user behavior result, the intention classification of wherein said user behavior result belonging to described historical search statement is determined;
Using user behavior result described in described user behavior characteristic sum as training example, based on machine learning algorithm, set up the personal behavior model corresponding with the intention classification belonging to described historical search statement.
4. method according to claim 2, is characterized in that, the intention classification belonging to described historical search statement and described historical behavior data, sets up personal behavior model, comprising:
From described historical search behavioral data, be extracted in user behavior feature corresponding to historical search result under described historical search statement and user behavior result, the intention classification of wherein said user behavior result belonging to described historical search statement is determined;
Using the intention category feature belonging to described user behavior feature, described historical search statement and described user behavior result as training example, set up personal behavior model based on machine learning algorithm.
5. the method according to any one of claim 1-4, is characterized in that, intention classification comprises following at least one: addressing class intention, info class intention and commodity class intention.
6. a searcher, is characterized in that, comprising:
Intention classification recognition unit, for identifying the intention classification belonging to the target search statement that active user inputs;
User's behavior prediction unit, for the intention classification obtained according to the identification of described intention classification recognition unit and the personal behavior model be pre-created, active user is to the behavior of candidate search result in prediction, and the intention classification of wherein said personal behavior model belonging to the historical search statement of historic user is determined;
Search result output unit, for predicting the outcome of obtaining based on described user's behavior prediction unit, determines this Search Results exported from described candidate search result.
7. device according to claim 6, is characterized in that, also comprises personal behavior model and sets up unit, and wherein said personal behavior model is set up unit and comprised:
Data acquisition subelement, for obtaining the historical behavior data of historic user;
Intention assessment subelement, for identifying the intention classification belonging to the historical search statement in described historical behavior data;
Subelement set up by model, for the intention classification belonging to described historical search statement and described historical behavior data, sets up personal behavior model.
8. device according to claim 7, is characterized in that, subelement set up by described model, specifically for:
From described historical behavior data, be extracted in user behavior feature corresponding to historical search result under described historical search statement and user behavior result, the intention classification of wherein said user behavior result belonging to described historical search statement is determined;
Using user behavior result described in described user behavior characteristic sum as training example, based on machine learning algorithm, set up the personal behavior model corresponding with the intention classification belonging to described historical search statement.
9. device according to claim 7, is characterized in that, subelement set up by described model, specifically for:
From described historical search behavioral data, be extracted in user behavior feature corresponding to historical search result under described historical search statement and user behavior result, the intention classification of wherein said user behavior result belonging to described historical search statement is determined;
Using the intention category feature belonging to described user behavior feature, described historical search statement and described user behavior result as training example, set up personal behavior model based on machine learning algorithm.
10. the method according to any one of claim 6-9, is characterized in that, intention classification comprises following at least one: addressing class intention, info class intention and commodity class intention.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510463375.6A CN105069077A (en) | 2015-07-31 | 2015-07-31 | Search method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510463375.6A CN105069077A (en) | 2015-07-31 | 2015-07-31 | Search method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105069077A true CN105069077A (en) | 2015-11-18 |
Family
ID=54498447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510463375.6A Pending CN105069077A (en) | 2015-07-31 | 2015-07-31 | Search method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105069077A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874467A (en) * | 2017-02-15 | 2017-06-20 | 百度在线网络技术（北京）有限公司 | Method and apparatus for providing Search Results |
| CN106951433A (en) * | 2017-02-04 | 2017-07-14 | 百度在线网络技术（北京）有限公司 | A kind of search method and device |
| CN107194515A (en) * | 2017-05-27 | 2017-09-22 | 冯小平 | Determine user's current behavior and the method and apparatus for predicting user view |
| CN107274217A (en) * | 2017-05-27 | 2017-10-20 | 冯小平 | Determine user's current behavior and the method and apparatus for predicting user view |
| CN107357875A (en) * | 2017-07-04 | 2017-11-17 | 北京奇艺世纪科技有限公司 | A kind of voice search method, device and electronic equipment |
| CN110968686A (en) * | 2018-09-28 | 2020-04-07 | 百度在线网络技术（北京）有限公司 | Intention recognition method, device, equipment and computer readable medium |
| CN111798941A (en) * | 2019-04-04 | 2020-10-20 | Iqvia 有限公司 | Predictive system for generating clinical queries |
| WO2021057741A1 (en) * | 2019-09-23 | 2021-04-01 | 杭州海康威视数字技术股份有限公司 | Image search method and device, and server and storage medium |
| CN112699277A (en) * | 2019-10-23 | 2021-04-23 | 阿里巴巴集团控股有限公司 | Data detection method and device |
| CN113343028A (en) * | 2021-05-31 | 2021-09-03 | 北京达佳互联信息技术有限公司 | Method and device for training intention determination model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060224587A1 (en) * | 2005-03-31 | 2006-10-05 | Google, Inc. | Systems and methods for modifying search results based on a user's history |
| CN102760138A (en) * | 2011-04-26 | 2012-10-31 | 北京百度网讯科技有限公司 | Classification method and device for user network behaviors and search method and device for user network behaviors |
| CN103020289A (en) * | 2012-12-25 | 2013-04-03 | 浙江鸿程计算机系统有限公司 | Method for providing individual needs of search engine user based on log mining |
| CN103646070A (en) * | 2013-12-06 | 2014-03-19 | 北京趣拿软件科技有限公司 | Data processing method and device for search engine |
| CN104462293A (en) * | 2014-11-27 | 2015-03-25 | 百度在线网络技术（北京）有限公司 | Search processing method and method and device for generating search result ranking model |
-
2015
- 2015-07-31 CN CN201510463375.6A patent/CN105069077A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060224587A1 (en) * | 2005-03-31 | 2006-10-05 | Google, Inc. | Systems and methods for modifying search results based on a user's history |
| CN102760138A (en) * | 2011-04-26 | 2012-10-31 | 北京百度网讯科技有限公司 | Classification method and device for user network behaviors and search method and device for user network behaviors |
| CN103020289A (en) * | 2012-12-25 | 2013-04-03 | 浙江鸿程计算机系统有限公司 | Method for providing individual needs of search engine user based on log mining |
| CN103646070A (en) * | 2013-12-06 | 2014-03-19 | 北京趣拿软件科技有限公司 | Data processing method and device for search engine |
| CN104462293A (en) * | 2014-11-27 | 2015-03-25 | 百度在线网络技术（北京）有限公司 | Search processing method and method and device for generating search result ranking model |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106951433A (en) * | 2017-02-04 | 2017-07-14 | 百度在线网络技术（北京）有限公司 | A kind of search method and device |
| CN106951433B (en) * | 2017-02-04 | 2020-08-28 | 百度在线网络技术（北京）有限公司 | Retrieval method and device |
| CN106874467B (en) * | 2017-02-15 | 2019-12-06 | 百度在线网络技术（北京）有限公司 | Method and apparatus for providing search results |
| US11669579B2 (en) | 2017-02-15 | 2023-06-06 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and apparatus for providing search results |
| CN106874467A (en) * | 2017-02-15 | 2017-06-20 | 百度在线网络技术（北京）有限公司 | Method and apparatus for providing Search Results |
| WO2018149115A1 (en) * | 2017-02-15 | 2018-08-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for providing search results |
| CN107274217A (en) * | 2017-05-27 | 2017-10-20 | 冯小平 | Determine user's current behavior and the method and apparatus for predicting user view |
| CN107194515A (en) * | 2017-05-27 | 2017-09-22 | 冯小平 | Determine user's current behavior and the method and apparatus for predicting user view |
| CN107357875A (en) * | 2017-07-04 | 2017-11-17 | 北京奇艺世纪科技有限公司 | A kind of voice search method, device and electronic equipment |
| CN110968686A (en) * | 2018-09-28 | 2020-04-07 | 百度在线网络技术（北京）有限公司 | Intention recognition method, device, equipment and computer readable medium |
| CN111798941A (en) * | 2019-04-04 | 2020-10-20 | Iqvia 有限公司 | Predictive system for generating clinical queries |
| CN111798941B (en) * | 2019-04-04 | 2023-10-13 | Iqvia 有限公司 | Predictive system for generating clinical queries |
| WO2021057741A1 (en) * | 2019-09-23 | 2021-04-01 | 杭州海康威视数字技术股份有限公司 | Image search method and device, and server and storage medium |
| CN112699277A (en) * | 2019-10-23 | 2021-04-23 | 阿里巴巴集团控股有限公司 | Data detection method and device |
| CN113343028A (en) * | 2021-05-31 | 2021-09-03 | 北京达佳互联信息技术有限公司 | Method and device for training intention determination model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105069077A (en) | Search method and device | |
| JP7343568B2 (en) | Identifying and applying hyperparameters for machine learning | |
| CN105247507B (en) | Method, system and storage medium for the influence power score for determining brand | |
| CN102760138B (en) | Classification method and device for user network behaviors and search method and device for user network behaviors | |
| CN102073725B (en) | Method for searching structured data and search engine system for implementing same | |
| CN102053983B (en) | Method, system and device for querying vertical search | |
| CN110909182B (en) | Multimedia resource searching method, device, computer equipment and storage medium | |
| CN108280114B (en) | Deep learning-based user literature reading interest analysis method | |
| CN105760400B (en) | A kind of PUSH message sort method and device based on search behavior | |
| CN105160545B (en) | Method and device for determining release information style | |
| CN106919625B (en) | Internet user attribute identification method and device | |
| CN105677780A (en) | Scalable user intent mining method and system thereof | |
| CN106251174A (en) | Information recommendation method and device | |
| CN105224554A (en) | Search word is recommended to carry out method, system, server and the intelligent terminal searched for | |
| CN110852793A (en) | Document recommendation method and device and electronic equipment | |
| JP2013531289A (en) | Use of model information group in search | |
| CN105045901A (en) | Search keyword push method and device | |
| CN111444334A (en) | Data processing method, text recognition device and computer equipment | |
| CN104216881A (en) | Method and device for recommending individual labels | |
| CN105740268A (en) | Information pushing method and apparatus | |
| CN104915426B (en) | Information sorting method, the method and device for generating information sorting model | |
| CN110046298A (en) | Query word recommendation method and device, terminal device and computer readable medium | |
| CN110134845A (en) | Project public sentiment monitoring method, device, computer equipment and storage medium | |
| CN108959580A (en) | A kind of optimization method and system of label data | |
| CN103970858A (en) | Recommended content determining system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151118 |
|
| RJ01 | Rejection of invention patent application after publication |