CN105335519A - Model generation method and device as well as recommendation method and device - Google Patents
Model generation method and device as well as recommendation method and device Download PDFInfo
- Publication number
- CN105335519A CN105335519A CN201510794561.8A CN201510794561A CN105335519A CN 105335519 A CN105335519 A CN 105335519A CN 201510794561 A CN201510794561 A CN 201510794561A CN 105335519 A CN105335519 A CN 105335519A
- Authority
- CN
- China
- Prior art keywords
- entity
- candidate
- vector
- degree
- designated entities
- 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
-
- 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 provides a model generation method, a model generation device, a recommendation method and a recommendation device. The model generation method in the embodiment of the invention comprises the following steps: acquiring a document content feature vector of each entity in a mapping knowledge domain, logical relationship feature vectors among the entities, a user behavior relationship feature vector of each entity and at least one of feature vectors of each entity, then performing machine learning according to the document content feature vectors, the logical relationship feature vectors, the user behavior relationship feature vectors and the at least one of the feature vectors, and generating a deep fusion model. Therefore, according to the technical scheme provided by the embodiment of the invention, the deep fusion model can be generated by integrating various relationships among the entities and can be used for acquiring the surprise degree among the entities, so that the entities can be recommended to users based on the surprise degree, the search recommendation requirements of the users are met, and the click rate of the recommended entities is improved.
Description
[technical field]
The present invention relates to search technique field, particularly relate to a kind of model generating method and device, recommend method and device.
[background technology]
At present, when carrying out search and recommending, when the main search demand being based on user is met, by providing other contents that may be interested in relevant to query word to user, the potential demand of user is excited.Such as, please refer to Fig. 1, it is that in prior art, knowledge based collection of illustrative plates carries out searching for first exemplary plot of recommending, as shown in the figure, when user's inquiry " Princeton University ", can recommend the famous alumnus of the Princeton University shown in Fig. 1 in the non-search results area of search results pages, this is the recommended entity very relevant to query word " Princeton University ".
But in prior art, when knowledge based collection of illustrative plates carries out search recommendation, the entity of recommendation is well-known often, the user interest that can not cause.Therefore, this search way of recommendation can not meet user search recommended requirements, causes the clicking rate of recommended entity lower.
[summary of the invention]
In view of this, embodiments provide a kind of model generating method and device, recommend method and device, degree of depth Fusion Model is generated by the various relations between integral entity, degree of depth Fusion Model may be used for obtaining the pleasantly surprised degree between entity, thus can based on pleasantly surprised degree to user's recommended entity, meet the search recommended requirements of user, improve the clicking rate of recommended entity.
The one side of the embodiment of the present invention, provides a kind of model generating method, comprising:
In the proper vector of each entity of user behavior relationship characteristic vector sum of the logic association relationship characteristic vector in acquire knowledge collection of illustrative plates between the document content proper vector of each entity, each entity, each entity at least one;
In proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, at least one carries out machine learning, generates degree of depth Fusion Model.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, obtains the user behavior relationship characteristic vector of each entity, comprising:
Obtain the historical search behavior record of user;
According to described historical search behavior record, obtain user for the search behavior of described each entity and the behavior of click;
According to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, obtain the proper vector of each entity, comprising: according to the entity defined in knowledge mapping, is proper vector described in each entity stochastic generation.
The one side of the embodiment of the present invention, provides a kind of recommend method, comprising:
Obtain candidate's entity that designated entities is corresponding;
By the document content proper vector of described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize a described method in claim 1 to 3 to obtain;
According to described pleasantly surprised degree and described candidate's entity, obtain the recommended entity that described designated entities is corresponding.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, and candidate's entity that described acquisition designated entities is corresponding comprises:
Vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity; Or, according to the entity defined in knowledge mapping, obtain described candidate's entity.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding, comprising:
According to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results, and will sort in ranking results at least one candidate's entity forward as recommended entity corresponding to described designated entities.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described method also comprises:
The recommended entity corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the proper vector obtained after adjustment is for generating described degree of depth Fusion Model.
The one side of the embodiment of the present invention, provides a kind of model generation device, comprising:
Vector acquiring unit, between the document content proper vector of entity each in acquire knowledge collection of illustrative plates, each entity logic association relationship characteristic vector, each entity each entity of user behavior relationship characteristic vector sum proper vector at least one;
Model generation unit, at least one carries out machine learning in proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, generates degree of depth Fusion Model.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described vectorial acquiring unit, specifically for:
Obtain the historical search behavior record of user;
According to described historical search behavior record, obtain user for the search behavior of described each entity and the behavior of click;
According to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described vectorial acquiring unit, specifically for: according to the entity defined in knowledge mapping, be proper vector described in each entity stochastic generation.
The one side of the embodiment of the present invention, provides a kind of recommendation apparatus, comprising:
Entity acquiring unit, for obtaining candidate's entity corresponding to designated entities;
Pleasantly surprised degree acquiring unit, for the document content proper vector by described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize a described device in claim 8 to 10 to generate;
Entity handles unit, for according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described entity acquiring unit, specifically for:
Vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity; Or, according to the entity defined in knowledge mapping, obtain described candidate's entity.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described entity handles unit, specifically for:
According to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results, and will sort in ranking results at least one candidate's entity forward as recommended entity corresponding to described designated entities.
Aspect as above and arbitrary possible implementation, provide a kind of implementation further, described device also comprises:
Vector adjustment unit, for the recommended entity corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the proper vector obtained after adjustment is for generating described degree of depth Fusion Model.
As can be seen from the above technical solutions, the embodiment of the present invention has following beneficial effect:
The technical scheme that the embodiment of the present invention provides can generate degree of depth Fusion Model by the various relations between integral entity, and this degree of depth Fusion Model may be used for obtaining the pleasantly surprised degree between entity, thus can based on pleasantly surprised degree to user's recommended entity.With prior art, only knowledge based collection of illustrative plates carry out search for recommend mode compare, the recommended entity that the embodiment of the present invention provides can cause the interest of user more, so can meet the search recommended requirements of user, improves the clicking rate of recommended entity.
[accompanying drawing explanation]
In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, be briefly described to the accompanying drawing used required in embodiment below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is that in prior art, knowledge based collection of illustrative plates carries out searching for first exemplary plot of recommending;
Fig. 2 is the schematic flow sheet of the model generating method that the embodiment of the present invention provides;
Fig. 3 is the generation exemplary plot of the degree of depth Fusion Model that the embodiment of the present invention provides;
Fig. 4 is the schematic flow sheet of the recommend method that the embodiment of the present invention provides;
Fig. 5 is first exemplary plot of carrying out searching for recommendation based on degree of depth Fusion Model that the embodiment of the present invention provides;
Fig. 6 is that in prior art, knowledge based collection of illustrative plates carries out searching for second exemplary plot of recommending;
Fig. 7 is second exemplary plot of carrying out searching for recommendation based on degree of depth Fusion Model that the embodiment of the present invention provides;
Fig. 8 is the functional block diagram of the model generation device that the embodiment of the present invention provides;
Fig. 9 is the functional block diagram of the recommendation apparatus that the embodiment of the present invention provides.
[embodiment]
Technical scheme for a better understanding of the present invention, is described in detail the embodiment of the present invention below in conjunction with accompanying drawing.
Should be clear and definite, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making other embodiments all obtained under creative work prerequisite, belong to the scope of protection of the invention.
The term used in embodiments of the present invention is only for the object describing specific embodiment, and not intended to be limiting the present invention." one ", " described " and " being somebody's turn to do " of the singulative used in the embodiment of the present invention and appended claims is also intended to comprise most form, unless context clearly represents other implications.
Should be appreciated that term "and/or" used herein is only a kind of incidence relation describing affiliated partner, can there are three kinds of relations in expression, and such as, A and/or B, can represent: individualism A, exists A and B simultaneously, these three kinds of situations of individualism B.In addition, character "/" herein, general expression forward-backward correlation is to the relation liking a kind of "or".
Depend on linguistic context, word as used in this " if " can be construed as into " ... time " or " when ... time " or " in response to determining " or " in response to detection ".Similarly, depend on linguistic context, phrase " if determination " or " if detecting (the conditioned disjunction event of statement) " can be construed as " when determining " or " in response to determining " or " when detecting (the conditioned disjunction event of statement) " or " in response to detection (the conditioned disjunction event of statement) ".
Embodiment one
The embodiment of the present invention provides a kind of model generating method, please refer to Fig. 2, the schematic flow sheet of its model generating method provided for the embodiment of the present invention, and as shown in the figure, the method comprises the following steps:
S201, in the proper vector of each entity of user behavior relationship characteristic vector sum of the logic association relationship characteristic vector in acquire knowledge collection of illustrative plates between the document content proper vector of each entity, each entity, each entity at least one.
S202, in proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, at least one carries out machine learning, generates degree of depth Fusion Model.
It should be noted that, in knowledge mapping, define the relevant information of each entity and each entity; Described entity refers to real-life things, as personage, article, virtual portrait or place etc.
Please refer to Fig. 3, the generation exemplary plot of its degree of depth Fusion Model provided for the embodiment of the present invention, as shown in the figure, in the embodiment of the present invention, before generation degree of depth Fusion Model, for two entity e1 and e2 any in knowledge mapping, need to obtain document content proper vector s1, the document content proper vector s2 of entity e2 of entity e1, user behavior relationship characteristic vector c, the proper vector p1 of entity e1 and the proper vector p2 of entity e2 of logic association relationship characteristic vector k, entity e1 and the entity e2 between entity e1 and entity e2.
Illustrate, in the embodiment of the present invention, the method obtaining the document content proper vector of each entity can include but not limited to: can utilize convolutional neural networks, modeling is carried out to the document d1 of entity e1, and modeling is carried out to the document d2 of entity e2, to obtain the document content proper vector s1 of entity e1, and the document content proper vector s2 of entity e2.
Such as, be described for entity e1: the document d1 that first can obtain entity e1 from knowledge mapping, such as, the text in the encyclopaedia page of entity e1 can as the document d1 of entity e1.Then, from document d1, word feature vector w1 ~ wn is extracted.Then, word feature vector w1 ~ wn carries out convolution algorithm in convolutional layer, to obtain vector characteristics.Finally, at maximum pond layer, the process of maximal value pondization is carried out, to obtain the document content proper vector s1 of entity e1 to the vector characteristics that convolutional layer exports.Wherein, the document content proper vector of entity with the convolution model used during convolution algorithm, when can generate degree of depth Fusion Model, carry out based on deep neural network carrying out Automatic Optimal in the process of degree of depth machine training.
Illustrate, in the embodiment of the present invention, the method obtaining the logic association relationship characteristic vector between each entity can include but not limited to: can obtain the logic association relationship characteristic vector k between entity e1 and entity e2 from knowledge mapping.Be understandable that, logic association relationship characteristic vector k can represent the logic association relation in knowledge mapping between entity e1 and entity e2.
Illustrate, in the embodiment of the present invention, the method obtaining user behavior relation vector can include but not limited to:
First, the historical search behavior record of user is obtained.Then, according to the historical search behavior record of described user, obtain user for the search behavior of described each entity and the behavior of click.Finally, according to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
Be understandable that, in the embodiment of the present invention, some numerical value is comprised in the user behavior relationship characteristic vector of each entity, each numerical value can represent a kind of user behavior relation between this entity and another entity, so user behavior relationship characteristic vector also can be understood as the user behavior relationship characteristic vector between an entity and another entity.
Such as, after user has searched for entity e1 in a search engine, click the entity e2 in the recommended entity of search results pages right side of face, then the middle number of clicks of the user behavior relationship characteristic vector of entity e1 and entity e2 adds 1.And, after user has searched for entity e1 in a search engine, searching entities e2 in a search engine again, then the searching times in the user behavior relationship characteristic vector of entity e1 and entity e2 adds 1; And after user has searched for entity e1 in a search engine, comprise the information of another entity e2 in the Search Results clicked, then the number of hops in the user behavior relationship characteristic vector of entity e1 and entity e2 adds 1.Be understandable that, in above-mentioned acquisition user behavior relationship characteristic vector, the statistical of numerical value is only and illustrates, in the embodiment of the present invention, for according to user for described each entity search behavior and click behavior, the mode obtaining the user behavior relationship characteristic vector of described each entity is not particularly limited.
In the embodiment of the present invention, what comprise in the proper vector of entity is lower with the correlativity of this entity but other entities that user more can be caused pleasantly surprised, incidence relation not obvious especially with this entity.
Illustrate, in the embodiment of the present invention, the method obtaining the proper vector of each entity can include but not limited to:
Can, according to the entity defined in knowledge mapping, be proper vector described in each entity stochastic generation.Or, can also in the proper vector according to stochastic generation, after generating degree of depth Fusion Model, utilize degree of depth Fusion Model for the recommended entity of designated entities acquisition correspondence, then corresponding according to described designated entities recommended entity, adjust the proper vector of described designated entities, then, the proper vector obtained after recycling adjustment, re-starts machine learning, to generate new degree of depth Fusion Model, thus achieve continuing to optimize of proper vector and degree of depth Fusion Model.Or, can also carrying out in the process of degree of depth machine learning utilizing deep neural network, by the backpropagation Optimization Mechanism of training error, adjusting the proper vector of described designated entities.
In a concrete implementation procedure, as shown in Figure 3, can by least one input deep neural network in proper vector described in the described document content proper vector obtained, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, deep neural network carries out degree of depth machine learning to user preference, to generate degree of depth Fusion Model according to the vector of input.
Embodiment two
The embodiment of the present invention provides a kind of recommend method, and the degree of depth Fusion Model used in the recommend method that the present embodiment provides is the degree of depth Fusion Model generated in the model generating method utilizing above-described embodiment one to provide.Please refer to Fig. 4, the schematic flow sheet of its recommend method provided for the embodiment of the present invention, as shown in the figure, the method comprises the following steps:
S401, obtains candidate's entity that designated entities is corresponding.
S402, by the document content proper vector of described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize above-mentioned model generating method to obtain.
S403, according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding.
Illustrate, in the embodiment of the present invention, the method obtaining candidate's entity corresponding to designated entities can include but not limited to following two kinds:
The first: according to the entity defined in knowledge mapping, obtain described candidate's entity.Such as, can using all entities of defining in knowledge mapping all as described candidate's entity.
The second: according to the title of the designated entities of user's input, obtains some user behavior relationship characteristic vectors of designated entities; Then, vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity.
Be understandable that, can store it after the user behavior relationship characteristic vector obtaining each entity in above-described embodiment one, like this, when needs utilize degree of depth Fusion Model to obtain recommended entity, can according to the title of designated entities, the some user behavior relationship characteristics vector that finds designated entities corresponding, the user behavior relation in these user behavior relationship characteristic vector representations designated entities and multiple entity between each entity.In this method, can using the candidate entity corresponding as designated entities with the multiple entities that there is user behavior relationship characteristic vector between designated entities.
Compared with first method, make use of user behavior relationship characteristic vector in second method to screen the entity defined in knowledge mapping, to reduce the scope of candidate's entity, decrease calculated amount when utilizing degree of depth Fusion Model acquisition recommended entity, thus improve the efficiency that degree of depth Fusion Model obtains recommended entity.
In the embodiment of the present invention, when obtaining each candidate's entity pleasantly surprised and spending, can by the document content proper vector of described designated entities, logic association relationship characteristic vector between designated entities and candidate's entity, in the proper vector of the user behavior relationship characteristic vector sum designated entities of described designated entities at least one, and, in the document content proper vector of described candidate's entity and the proper vector of described candidate's entity at least one, input degree of depth Fusion Model, degree of depth Fusion Model can calculate and export the pleasantly surprised degree of each candidate's entity, thus obtain the pleasantly surprised degree of described candidate's entity.
It should be noted that, here the vector inputting degree of depth Fusion Model needs with when generating degree of depth Fusion Model, the vector used when carrying out degree of depth machine learning is consistent, such as, when generating degree of depth Fusion Model, if use the logic association relationship characteristic vector between designated entities and candidate's entity to carry out degree of depth machine learning, then need the logic association relationship characteristic vector between designated entities and candidate's entity to input degree of depth Fusion Model here.Or, again such as, when generating degree of depth Fusion Model, if use the proper vector of entity to carry out degree of depth machine learning, then need the proper vector of the proper vector of designated entities and candidate's entity to input degree of depth Fusion Model here.
Be understandable that, expected degree refers in recommendation results, and the recommended entity deriving from knowledge mapping and rule generation accounts for the ratio of all recommended entity.Pleasantly surprised degree equals 1 and deducts expected degree, pleasantly surprised degree refers in recommendation results, other entities outside the recommended entity deriving from knowledge mapping and generate rule account for the ratio of all recommended entity, for prediction after user inputs the title of designated entities, when providing recommended entity to user, user is for the pleasantly surprised degree of recommended entity.
Illustrate, according to described pleasantly surprised degree and described candidate's entity, the method obtaining recommended entity corresponding to described first instance can include but not limited to:
First, according to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results.Then, according to the recommendation number preset, extract in ranking results at least one the candidate's entity of the forward respective number that sorts, using at least one candidate's entity of extracting as recommended entity corresponding to described designated entities.
Be understandable that, can when exporting to user the Search Results matched with designated entities, the recommended entity of acquisition is recommended user, and such as, recommended entity can be presented in the right side of search results pages.
Optionally, in one of the present embodiment possible implementation, recommended entity that can also be corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the entity comprised in the proper vector of designated entities can be one or more in described recommended entity.Further, the proper vector obtained after adjustment may be used for carrying out degree of depth machine learning, to generate new degree of depth Fusion Model, new degree of depth Fusion Model can also be used for obtaining recommended entity further, by that analogy, so repeatedly constantly can be optimized adjustment to the proper vector of entity, and adjustment is optimized to degree of depth Fusion Model, thus improve constantly the acquisition accuracy of recommended entity, improve constantly the satisfaction of user to recommended entity, improve the clicking rate of recommended entity.
Such as, please refer to Fig. 5, what it provided for the embodiment of the present invention carries out searching for first exemplary plot of recommending based on degree of depth Fusion Model, as shown in Figure 5, if the name of the designated entities of user's input is called " Princeton University ", recommends if utilize knowledge based collection of illustrative plates in prior art to carry out search, will the recommended entity shown in Fig. 1 be obtained, these recommended entity are well-known for user, cannot cause the interest of user.But, the degree of depth Fusion Model utilizing the embodiment of the present invention to provide, the recommended entity shown in Fig. 5 can be obtained, these recommended entity are entities obviously not relevant to designated entities, but some scholar-tyrant, one who exercises autocratic control in academic and educational circles, obviously these scholar-tyrant, one who exercises autocratic control in academic and educational circles more can cause user interest, and then the click of activated user, excite the potential search need of user, therefore these recommended entity more can be met consumers' demand, and improve the clicking rate of recommending accuracy rate and recommended entity.
Or, again such as, please refer to Fig. 6 and Fig. 7, be respectively knowledge based collection of illustrative plates in prior art and carry out searching for second exemplary plot of recommending, and what the embodiment of the present invention provided carries out searching for second exemplary plot of recommending based on degree of depth Fusion Model.
When user search " All Saints' Day ", the potential demand of user can comprise: the film that the film of terrible terror, All Saints' Day are correlated with, game/theme, other fearful ghost/monster/biologies of preparing stage property needed for the All Saints' Day, All Saints' Day party.As shown in Figure 6, the recommended entity shown to user is the Chinese and western red-letter day of being correlated with the All Saints' Day by knowledge based collection of illustrative plates, and the pleasantly surprised degree brought to user is lower.But, as shown in Figure 7, if the recommended entity utilizing degree of depth Fusion Model to generate contains 7 terrified films (entity that in Fig. 7, dotted line frame identifies) and contains whole 5 other recommended entity (realizing the entity that frame identifies in Fig. 7), the coverage rate of these recommended entity is wider, and the pleasantly surprised degree that the recommended entity shown in Fig. 7 is brought to user is larger.In Fig. 7, the recommended entity shown in solid box and dotted line collimation mark represents the entity that in experiment, user's clicking rate is high.Can find out, the pleasantly surprised degree of the entity that degree of depth Fusion Model is excavated obtains more concern and the interest of user really.
The embodiment of the present invention provides the device embodiment realizing each step and method in said method embodiment further.
Please refer to Fig. 8, the functional block diagram of its model generation device provided for the embodiment of the present invention.As shown in the figure, this device comprises:
Vector acquiring unit 81, between the document content proper vector of entity each in acquire knowledge collection of illustrative plates, each entity logic association relationship characteristic vector, each entity each entity of user behavior relationship characteristic vector sum proper vector at least one;
Model generation unit 82, at least one carries out machine learning in proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, generates degree of depth Fusion Model.
In a concrete implementation procedure, described vectorial acquiring unit 81, specifically for:
Obtain the historical search behavior record of user;
According to described historical search behavior record, obtain user for the search behavior of described each entity and the behavior of click;
According to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
In a concrete implementation procedure, described vectorial acquiring unit 81, specifically for: according to the entity defined in knowledge mapping, be proper vector described in each entity stochastic generation.
Because each unit in the present embodiment can perform the method shown in Fig. 2, the part that the present embodiment is not described in detail, can with reference to the related description to Fig. 2.
Please refer to Fig. 9, the functional block diagram of its recommendation apparatus provided for the embodiment of the present invention.As shown in the figure, this device comprises:
Entity acquiring unit 91, for obtaining candidate's entity corresponding to designated entities;
Pleasantly surprised degree acquiring unit 92, for the document content proper vector by described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize model generation device to generate;
Entity handles unit 93, for according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding.
In a concrete implementation procedure, described entity acquiring unit 91, specifically for:
Vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity; Or, according to the entity defined in knowledge mapping, obtain described candidate's entity.
In a concrete implementation procedure, described entity handles unit 93, specifically for:
According to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results, and will sort in ranking results at least one candidate's entity forward as recommended entity corresponding to described designated entities.
Optionally, in one of the present embodiment possible implementation, described device also comprises:
Vector adjustment unit 94, for the recommended entity corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the proper vector obtained after adjustment is for generating described degree of depth Fusion Model.
Because each unit in the present embodiment can perform the method shown in Fig. 4, the part that the present embodiment is not described in detail, can with reference to the related description to Fig. 4.
The technical scheme of the embodiment of the present invention has following beneficial effect:
In the embodiment of the present invention, by logic association relation vector, the user behavior relation vector of each entity and the proper vector of each entity between the document content vector of entity each in acquire knowledge collection of illustrative plates, each entity; Thus, carry out machine learning according to described document content vector, described logic association relation vector, described user behavior relation vector and described proper vector, generate degree of depth Fusion Model.
The technical scheme that the embodiment of the present invention provides can generate degree of depth Fusion Model by the various relations between integral entity, and this degree of depth Fusion Model may be used for obtaining the pleasantly surprised degree between entity, thus can based on pleasantly surprised degree to user's recommended entity.With prior art, only knowledge based collection of illustrative plates carry out search for recommend mode compare, the recommended entity that the embodiment of the present invention provides can cause the interest of user more, so can meet the search recommended requirements of user, improves the clicking rate of recommended entity.
Those skilled in the art can be well understood to, and for convenience and simplicity of description, the system of foregoing description, the specific works process of device and unit, with reference to the corresponding process in preceding method embodiment, can not repeat them here.
In several embodiment provided by the present invention, should be understood that, disclosed system, apparatus and method, can realize by another way.Such as, device embodiment described above is only schematic, such as, the division of described unit, is only a kind of logic function and divides, and actual can have other dividing mode when realizing, such as, multiple unit or assembly can in conjunction with or another system can be integrated into, or some features can be ignored, or do not perform.Another point, shown or discussed coupling each other or direct-coupling or communication connection can be by some interfaces, and the indirect coupling of device or unit or communication connection can be electrical, machinery or other form.
The described unit illustrated as separating component or can may not be and physically separates, and the parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of unit wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, also can be that the independent physics of unit exists, also can two or more unit in a unit integrated.Above-mentioned integrated unit both can adopt the form of hardware to realize, and the form that hardware also can be adopted to add SFU software functional unit realizes.
The above-mentioned integrated unit realized with the form of SFU software functional unit, can be stored in a computer read/write memory medium.Above-mentioned SFU software functional unit is stored in a storage medium, comprising some instructions in order to make a computer installation (can be personal computer, server, or network equipment etc.) or processor (Processor) perform the part steps of method described in each embodiment of the present invention.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (Read-OnlyMemory, ROM), random access memory (RandomAccessMemory, RAM), magnetic disc or CD etc. various can be program code stored medium.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within the scope of protection of the invention.
Claims (14)
1. a model generating method, is characterized in that, described method comprises:
In the proper vector of each entity of user behavior relationship characteristic vector sum of the logic association relationship characteristic vector in acquire knowledge collection of illustrative plates between the document content proper vector of each entity, each entity, each entity at least one;
In proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, at least one carries out machine learning, generates degree of depth Fusion Model.
2. method according to claim 1, is characterized in that, obtains the user behavior relationship characteristic vector of each entity, comprising:
Obtain the historical search behavior record of user;
According to described historical search behavior record, obtain user for the search behavior of described each entity and the behavior of click;
According to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
3. method according to claim 1, is characterized in that, obtains the proper vector of each entity, comprising: according to the entity defined in knowledge mapping, is proper vector described in each entity stochastic generation.
4. a recommend method, is characterized in that, described method comprises:
Obtain candidate's entity that designated entities is corresponding;
By the document content proper vector of described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize a described method in claim 1 to 3 to obtain;
According to described pleasantly surprised degree and described candidate's entity, obtain the recommended entity that described designated entities is corresponding.
5. method according to claim 4, is characterized in that, candidate's entity that described acquisition designated entities is corresponding, comprising:
Vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity; Or, according to the entity defined in knowledge mapping, obtain described candidate's entity.
6. method according to claim 4, is characterized in that, according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding, comprising:
According to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results, and will sort in ranking results at least one candidate's entity forward as recommended entity corresponding to described designated entities.
7. the method according to any one of claim 4 to 6, is characterized in that, described method also comprises:
The recommended entity corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the proper vector obtained after adjustment is for generating described degree of depth Fusion Model.
8. a model generation device, is characterized in that, described device comprises:
Vector acquiring unit, between the document content proper vector of entity each in acquire knowledge collection of illustrative plates, each entity logic association relationship characteristic vector, each entity each entity of user behavior relationship characteristic vector sum proper vector at least one;
Model generation unit, at least one carries out machine learning in proper vector according to described document content proper vector, described logic association relationship characteristic vector, described user behavior relationship characteristic vector sum, generates degree of depth Fusion Model.
9. device according to claim 8, is characterized in that, described vectorial acquiring unit, specifically for:
Obtain the historical search behavior record of user;
According to described historical search behavior record, obtain user for the search behavior of described each entity and the behavior of click;
According to user for the search behavior of described each entity and the behavior of click, obtain the user behavior relationship characteristic vector of described each entity.
10. device according to claim 8, is characterized in that, described vectorial acquiring unit, specifically for: according to the entity defined in knowledge mapping, be proper vector described in each entity stochastic generation.
11. 1 kinds of recommendation apparatus, is characterized in that, described device comprises:
Entity acquiring unit, for obtaining candidate's entity corresponding to designated entities;
Pleasantly surprised degree acquiring unit, for the document content proper vector by described designated entities, between designated entities and candidate's entity logic association relationship characteristic vector, described designated entities user behavior relationship characteristic vector sum designated entities character vector at least one, and in the proper vector of the document content proper vector of described candidate's entity and described candidate's entity at least one, input degree of depth Fusion Model, to obtain the pleasantly surprised degree of described candidate's entity; Described degree of depth Fusion Model is utilize a described device in claim 8 to 10 to generate;
Entity handles unit, for according to described pleasantly surprised degree and described candidate's entity, obtains the recommended entity that described designated entities is corresponding.
12. devices according to claim 11, is characterized in that, described entity acquiring unit, specifically for:
Vectorial according to the user behavior relationship characteristic of described designated entities and described designated entities, obtain described candidate's entity; Or, according to the entity defined in knowledge mapping, obtain described candidate's entity.
13. devices according to claim 11, is characterized in that, described entity handles unit, specifically for:
According to the order that described pleasantly surprised degree is descending, described candidate's entity is sorted, to obtain ranking results, and will sort in ranking results at least one candidate's entity forward as recommended entity corresponding to described designated entities.
14., according to claim 11 to the device according to any one of 13, is characterized in that, described device also comprises:
Vector adjustment unit, for the recommended entity corresponding according to described designated entities, adjusts the proper vector of described designated entities, and the proper vector obtained after adjustment is for generating described degree of depth Fusion Model.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510794561.8A CN105335519B (en) | 2015-11-18 | 2015-11-18 | Model generation method and device and recommendation method and device |
| PCT/CN2016/089648 WO2017084362A1 (en) | 2015-11-18 | 2016-07-11 | Model generation method, recommendation method and corresponding apparatuses, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510794561.8A CN105335519B (en) | 2015-11-18 | 2015-11-18 | Model generation method and device and recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105335519A true CN105335519A (en) | 2016-02-17 |
| CN105335519B CN105335519B (en) | 2021-08-17 |
Family
ID=55286046
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510794561.8A Active CN105335519B (en) | 2015-11-18 | 2015-11-18 | Model generation method and device and recommendation method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105335519B (en) |
| WO (1) | WO2017084362A1 (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105808931A (en) * | 2016-03-03 | 2016-07-27 | 北京大学深圳研究生院 | Knowledge graph based acupuncture and moxibustion decision support method and apparatus |
| CN105912500A (en) * | 2016-03-30 | 2016-08-31 | 百度在线网络技术（北京）有限公司 | Machine learning model generation method and machine learning model generation device |
| CN106022877A (en) * | 2016-05-19 | 2016-10-12 | 华南理工大学 | User mobile game behavior map-based game recommendation method |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| WO2017084362A1 (en) * | 2015-11-18 | 2017-05-26 | 百度在线网络技术（北京）有限公司 | Model generation method, recommendation method and corresponding apparatuses, device and storage medium |
| CN107330050A (en) * | 2017-06-28 | 2017-11-07 | 达而观信息科技（上海）有限公司 | A kind of personalized recommendation method and system |
| CN107463701A (en) * | 2017-08-15 | 2017-12-12 | 北京百度网讯科技有限公司 | Method and apparatus based on artificial intelligence pushed information stream |
| CN107506486A (en) * | 2017-09-21 | 2017-12-22 | 北京航空航天大学 | A kind of relation extending method based on entity link |
| CN108334588A (en) * | 2018-01-29 | 2018-07-27 | 北京搜狐新媒体信息技术有限公司 | A kind of user tag construction method and device |
| CN108345702A (en) * | 2018-04-10 | 2018-07-31 | 北京百度网讯科技有限公司 | Entity recommends method and apparatus |
| CN108509479A (en) * | 2017-12-13 | 2018-09-07 | 深圳市腾讯计算机系统有限公司 | Entity recommends method and device, terminal and readable storage medium storing program for executing |
| CN108804387A (en) * | 2017-04-27 | 2018-11-13 | 腾讯科技（深圳）有限公司 | Target user determines method and device |
| TWI651664B (en) * | 2017-11-15 | 2019-02-21 | 財團法人資訊工業策進會 | Model generation server and its model generation method |
| CN109388793A (en) * | 2017-08-03 | 2019-02-26 | 阿里巴巴集团控股有限公司 | Entity mask method, intension recognizing method and corresponding intrument, computer storage medium |
| CN109829116A (en) * | 2019-02-14 | 2019-05-31 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, server and computer readable storage medium |
| CN110674394A (en) * | 2019-08-20 | 2020-01-10 | 腾讯科技（深圳）有限公司 | Knowledge graph-based information recommendation method and device and storage medium |
| CN110717099A (en) * | 2019-09-25 | 2020-01-21 | 优地网络有限公司 | Method and terminal for recommending film |
| CN110728783A (en) * | 2019-08-31 | 2020-01-24 | 苏州浪潮智能科技有限公司 | Self-correction method, system and equipment of face recognition system |
| CN111782774A (en) * | 2019-04-03 | 2020-10-16 | 北京嘀嘀无限科技发展有限公司 | Question recommendation method and device |
| CN112182115A (en) * | 2020-09-10 | 2021-01-05 | 珠海格力电器股份有限公司 | Relationship display method and device, storage medium and electronic device |
| CN112291272A (en) * | 2020-12-24 | 2021-01-29 | 鹏城实验室 | Network threat detection method, device, equipment and computer readable storage medium |
| CN108829854B (en) * | 2018-06-21 | 2021-08-31 | 北京百度网讯科技有限公司 | Method, apparatus, device and computer-readable storage medium for generating article |
| CN113536137A (en) * | 2021-08-13 | 2021-10-22 | 北京字节跳动网络技术有限公司 | Information display method and device and computer storage medium |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665394B (en) * | 2018-05-07 | 2021-11-30 | 北京文华在线教育科技股份有限公司 | Self-adaptive learning method and system |
| CN109657069B (en) * | 2018-12-11 | 2021-03-23 | 北京百度网讯科技有限公司 | Knowledge graph generation method and device |
| CN109697050B (en) * | 2019-01-07 | 2021-04-27 | 浙江大学 | Demand description model design method facing E-commerce field based on knowledge graph |
| CN109828965B (en) * | 2019-01-09 | 2021-06-15 | 千城数智(北京)网络科技有限公司 | Data processing method and electronic equipment |
| CN110555753A (en) * | 2019-08-14 | 2019-12-10 | 中国平安人寿保险股份有限公司 | recommendation-based ranking control method and device, computer equipment and storage medium |
| CN110888990B (en) * | 2019-11-22 | 2024-04-12 | 深圳前海微众银行股份有限公司 | Text recommendation method, device, equipment and medium |
| CN111428053B (en) * | 2020-03-30 | 2023-10-20 | 西安交通大学 | Construction method of tax field-oriented knowledge graph |
| CN112950321A (en) * | 2021-03-10 | 2021-06-11 | 北京汇钧科技有限公司 | Article recommendation method and device |
| CN117787392A (en) * | 2024-02-23 | 2024-03-29 | 支付宝(杭州)信息技术有限公司 | Knowledge graph fusion method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090125549A1 (en) * | 2007-11-13 | 2009-05-14 | Nec (China) Co., Ltd. | Method and system for calculating competitiveness metric between objects |
| CN103593792A (en) * | 2013-11-13 | 2014-02-19 | 复旦大学 | Individual recommendation method and system based on Chinese knowledge mapping |
| CN104090963A (en) * | 2014-07-14 | 2014-10-08 | 百度在线网络技术（北京）有限公司 | Search information recommendation method and device |
| CN104346446A (en) * | 2014-10-27 | 2015-02-11 | 百度在线网络技术（北京）有限公司 | Paper associated information recommendation method and device based on mapping knowledge domain |
| CN104484459A (en) * | 2014-12-29 | 2015-04-01 | 北京奇虎科技有限公司 | Method and device for combining entities in knowledge map |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101114294A (en) * | 2007-08-22 | 2008-01-30 | 杭州经合易智控股有限公司 | Self-help intelligent uprightness searching method |
| CN104035917B (en) * | 2014-06-10 | 2017-07-07 | 复旦大学 | A kind of knowledge mapping management method and system based on semantic space mapping |
| 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 |
| CN105335519B (en) * | 2015-11-18 | 2021-08-17 | 百度在线网络技术（北京）有限公司 | Model generation method and device and recommendation method and device |
-
2015
- 2015-11-18 CN CN201510794561.8A patent/CN105335519B/en active Active
-
2016
- 2016-07-11 WO PCT/CN2016/089648 patent/WO2017084362A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090125549A1 (en) * | 2007-11-13 | 2009-05-14 | Nec (China) Co., Ltd. | Method and system for calculating competitiveness metric between objects |
| CN103593792A (en) * | 2013-11-13 | 2014-02-19 | 复旦大学 | Individual recommendation method and system based on Chinese knowledge mapping |
| CN104090963A (en) * | 2014-07-14 | 2014-10-08 | 百度在线网络技术（北京）有限公司 | Search information recommendation method and device |
| CN104346446A (en) * | 2014-10-27 | 2015-02-11 | 百度在线网络技术（北京）有限公司 | Paper associated information recommendation method and device based on mapping knowledge domain |
| CN104484459A (en) * | 2014-12-29 | 2015-04-01 | 北京奇虎科技有限公司 | Method and device for combining entities in knowledge map |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017084362A1 (en) * | 2015-11-18 | 2017-05-26 | 百度在线网络技术（北京）有限公司 | Model generation method, recommendation method and corresponding apparatuses, device and storage medium |
| CN105808931A (en) * | 2016-03-03 | 2016-07-27 | 北京大学深圳研究生院 | Knowledge graph based acupuncture and moxibustion decision support method and apparatus |
| CN105808931B (en) * | 2016-03-03 | 2019-05-07 | 北京大学深圳研究生院 | A kind of the acupuncture decision support method and device of knowledge based map |
| CN105912500A (en) * | 2016-03-30 | 2016-08-31 | 百度在线网络技术（北京）有限公司 | Machine learning model generation method and machine learning model generation device |
| US11531926B2 (en) | 2016-03-30 | 2022-12-20 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and apparatus for generating machine learning model by using distributed computing framework |
| CN106022877A (en) * | 2016-05-19 | 2016-10-12 | 华南理工大学 | User mobile game behavior map-based game recommendation method |
| CN106649774A (en) * | 2016-12-27 | 2017-05-10 | 北京百度网讯科技有限公司 | Artificial intelligence-based object pushing method and apparatus |
| CN108804387B (en) * | 2017-04-27 | 2021-07-23 | 腾讯科技（深圳）有限公司 | Target user determination method and device |
| CN108804387A (en) * | 2017-04-27 | 2018-11-13 | 腾讯科技（深圳）有限公司 | Target user determines method and device |
| CN107330050A (en) * | 2017-06-28 | 2017-11-07 | 达而观信息科技（上海）有限公司 | A kind of personalized recommendation method and system |
| CN109388793A (en) * | 2017-08-03 | 2019-02-26 | 阿里巴巴集团控股有限公司 | Entity mask method, intension recognizing method and corresponding intrument, computer storage medium |
| CN109388793B (en) * | 2017-08-03 | 2023-04-07 | 阿里巴巴集团控股有限公司 | Entity marking method, intention identification method, corresponding device and computer storage medium |
| CN107463701A (en) * | 2017-08-15 | 2017-12-12 | 北京百度网讯科技有限公司 | Method and apparatus based on artificial intelligence pushed information stream |
| CN107463701B (en) * | 2017-08-15 | 2020-09-08 | 北京百度网讯科技有限公司 | Method and device for pushing information stream based on artificial intelligence |
| CN107506486A (en) * | 2017-09-21 | 2017-12-22 | 北京航空航天大学 | A kind of relation extending method based on entity link |
| TWI651664B (en) * | 2017-11-15 | 2019-02-21 | 財團法人資訊工業策進會 | Model generation server and its model generation method |
| CN108509479B (en) * | 2017-12-13 | 2022-02-11 | 深圳市腾讯计算机系统有限公司 | Entity recommendation method and device, terminal and readable storage medium |
| CN108509479A (en) * | 2017-12-13 | 2018-09-07 | 深圳市腾讯计算机系统有限公司 | Entity recommends method and device, terminal and readable storage medium storing program for executing |
| CN108334588A (en) * | 2018-01-29 | 2018-07-27 | 北京搜狐新媒体信息技术有限公司 | A kind of user tag construction method and device |
| CN108334588B (en) * | 2018-01-29 | 2021-03-23 | 北京搜狐新媒体信息技术有限公司 | User tag construction method and device |
| CN108345702A (en) * | 2018-04-10 | 2018-07-31 | 北京百度网讯科技有限公司 | Entity recommends method and apparatus |
| CN108829854B (en) * | 2018-06-21 | 2021-08-31 | 北京百度网讯科技有限公司 | Method, apparatus, device and computer-readable storage medium for generating article |
| CN109829116A (en) * | 2019-02-14 | 2019-05-31 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, server and computer readable storage medium |
| CN109829116B (en) * | 2019-02-14 | 2021-07-30 | 北京达佳互联信息技术有限公司 | Content recommendation method and device, server and computer readable storage medium |
| CN111782774B (en) * | 2019-04-03 | 2024-04-19 | 北京嘀嘀无限科技发展有限公司 | Method and device for recommending problems |
| CN111782774A (en) * | 2019-04-03 | 2020-10-16 | 北京嘀嘀无限科技发展有限公司 | Question recommendation method and device |
| CN110674394A (en) * | 2019-08-20 | 2020-01-10 | 腾讯科技（深圳）有限公司 | Knowledge graph-based information recommendation method and device and storage medium |
| CN110674394B (en) * | 2019-08-20 | 2024-03-15 | 腾讯科技（深圳）有限公司 | Knowledge graph-based information recommendation method, device and storage medium |
| CN110728783A (en) * | 2019-08-31 | 2020-01-24 | 苏州浪潮智能科技有限公司 | Self-correction method, system and equipment of face recognition system |
| CN110717099A (en) * | 2019-09-25 | 2020-01-21 | 优地网络有限公司 | Method and terminal for recommending film |
| CN112182115A (en) * | 2020-09-10 | 2021-01-05 | 珠海格力电器股份有限公司 | Relationship display method and device, storage medium and electronic device |
| CN112291272B (en) * | 2020-12-24 | 2021-05-11 | 鹏城实验室 | Network threat detection method, device, equipment and computer readable storage medium |
| CN112291272A (en) * | 2020-12-24 | 2021-01-29 | 鹏城实验室 | Network threat detection method, device, equipment and computer readable storage medium |
| CN113536137A (en) * | 2021-08-13 | 2021-10-22 | 北京字节跳动网络技术有限公司 | Information display method and device and computer storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2017084362A1 (en) | 2017-05-26 |
| CN105335519B (en) | 2021-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105335519A (en) | Model generation method and device as well as recommendation method and device | |
| US11397772B2 (en) | Information search method, apparatus, and system | |
| US11550871B1 (en) | Processing structured documents using convolutional neural networks | |
| US20210019341A1 (en) | Implementing a software action based on machine interpretation of a language input | |
| US10783361B2 (en) | Predictive analysis of target behaviors utilizing RNN-based user embeddings | |
| EP3617952A1 (en) | Information search method, apparatus and system | |
| CN103164463B (en) | Method and device for recommending labels | |
| CN106776673B (en) | Multimedia document summarization | |
| US20150213361A1 (en) | Predicting interesting things and concepts in content | |
| CN106682152B (en) | Personalized message recommendation method | |
| CN110704739A (en) | Resource recommendation method and device and computer storage medium | |
| CN103714063B (en) | Data analysing method and its system | |
| CN105677780A (en) | Scalable user intent mining method and system thereof | |
| CN109766557A (en) | A kind of sentiment analysis method, apparatus, storage medium and terminal device | |
| CN104102719A (en) | Track information pushing method and device | |
| US20180300296A1 (en) | Document similarity analysis | |
| CN104487918B (en) | Character string prediction is carried out from buffering area | |
| CN102135983A (en) | Group dividing method and device based on network user behavior | |
| CN106126605B (en) | Short text classification method based on user portrait | |
| CN105512180A (en) | Search recommendation method and device | |
| CN106354867A (en) | Multimedia resource recommendation method and device | |
| US20160042370A1 (en) | Providing survey content recommendations | |
| CN114036398A (en) | Content recommendation and ranking model training method, device, equipment and storage medium | |
| CN111563198B (en) | Material recall method, device, equipment and storage medium | |
| US9720914B2 (en) | Navigational aid for electronic books and documents |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |