CN108776676A - Information recommendation method, device, computer-readable medium and electronic equipment - Google Patents

Abstract

The embodiment provides a kind of information recommendation method, device, computer-readable medium and electronic equipments.The information recommendation method includes：Obtain the information to be recommended to user；According to the feature of the user, used intended display scheme when determining to user recommendation described information；Recommend described information to the user based on the intended display scheme.The technical solution of the embodiment of the present invention, to user's recommendation information, can realize the personalized displaying of recommendation information based on the exhibition scheme that the feature with user matches, and then can improve the clicking rate of the information of recommendation.

Description

Information recommendation method, device, computer-readable medium and electronic equipment

Technical field

The present invention relates to technical field of data processing, in particular to a kind of information recommendation method, device, computer Readable medium and electronic equipment.

Background technology

Currently, in the related technology when by identical commending contents to user, the clip Text that shows all be it is identical, Therefore user may be detectable the place interested of recommendation, into without clicking viewing content full text, seriously affect The clicking rate of recommendation.

It should be noted that information is only used for reinforcing the reason of the background to the present invention disclosed in above-mentioned background technology part Solution, therefore may include the information not constituted to the prior art known to persons of ordinary skill in the art.

Invention content

The embodiment of the present invention is designed to provide a kind of information recommendation method, device, computer-readable medium and electronics Equipment, and then the personalized displaying of recommendation information is realized at least to a certain extent, to improve the clicking rate of recommendation information.

Other characteristics and advantages of the present invention will be apparent from by the following detailed description, or partially by the present invention Practice and acquistion.

One side according to the ... of the embodiment of the present invention provides a kind of information recommendation method, including：It obtains to be recommended to use The information at family；According to the feature of the user, used intended display scheme when determining to user recommendation described information； Recommend described information to the user based on the intended display scheme.

One side according to the ... of the embodiment of the present invention provides a kind of information recommending apparatus, including：First acquisition unit, For obtaining the information to be recommended to user；Processing unit determines and recommends to the user for the feature according to the user Used intended display scheme when described information；Recommendation unit is pushed away for being based on the intended display scheme to the user Recommend described information.

In some embodiments of the invention, aforementioned schemes are based on, the processing unit includes：Exhibition scheme generates single Member, for generating multiple candidate exhibition schemes according to described information；Second acquisition unit, for obtaining each candidate displaying The content characteristic of scheme；Exhibition scheme selecting unit is used for according to the content characteristic, from the multiple candidate exhibition scheme Select the candidate exhibition scheme to match with the feature of the user as the intended display scheme.

In some embodiments of the invention, aforementioned schemes are based on, the exhibition scheme generation unit includes：First extraction Unit, for extracting the content of text in described information；Abstract and title generation unit, for according to the content of text, life At the abstract and title of multiple candidates.

In some embodiments of the invention, aforementioned schemes are based on, the abstract and title generation unit are used for：For institute Multiple dimensions of content of text are stated, abstract corresponding with each dimension in the multiple dimension and title are generated, it is described to obtain The synopsis and title of multiple candidates.

In some embodiments of the invention, aforementioned schemes are based on, the exhibition scheme generation unit includes：Second extraction Unit, for extracting picture and/or video in described information；Picture generation unit, for according to the picture and/or regarding Frequently, multiple candidate pictures are generated.

In some embodiments of the invention, aforementioned schemes are based on, the picture generation unit includes：First generates list Member, for using the picture extracted from described information as the candidate picture；And/or second generation unit, for from described At least one key frame is extracted in video, using at least one key frame as the candidate picture.

In some embodiments of the invention, aforementioned schemes are based on, second generation unit includes：Cluster cell is used Clustering processing is carried out in the picture frame for including by the video, obtains at least one class；Picture frame selecting unit is used for from each The nearest picture frame of chosen distance cluster centre is as the key frame in the class.

In some embodiments of the invention, aforementioned schemes are based on, second generation unit further includes：Cutting unit, For being at least by the video slicing before the picture frame that the video is included by the cluster cell carries out clustering processing One camera lens picture, and delete the picture frame except at least one camera lens picture.

In some embodiments of the invention, aforementioned schemes are based on, the cutting unit includes：Difference computing unit is used In the difference for calculating adjacent two field pictures in the video successively；Cut-off determination unit, for according to adjacent in the video The difference of two field pictures determines the shot segmentation point in the video；Execution unit is used for based on determining shot segmentation point, It is at least one camera lens picture by the video slicing.

In some embodiments of the invention, aforementioned schemes are based on, the cut-off determination unit is used for：In the video In L-th frame image and L-1 frame images difference be greater than or equal to first threshold when, by the L-1 frames image with it is described As a shot segmentation point of the video between L-th frame image, wherein the L is greater than or equal to 2.

In some embodiments of the invention, aforementioned schemes are based on, the cut-off determination unit is additionally operable to：Described The difference of L frames image and the L-1 frame images is less than the first threshold, and when more than or equal to second threshold, from described The L-th frame image in video starts, the difference for the adjacent two field pictures that add up successively；When the L-th frame image is regarded with described The difference summation of adjacent two field pictures is greater than or equal to the first threshold, and the L+n between L+n frame images in frequency When frame image and the difference of the L+n-1 frame images in the video are less than the second threshold, by the L-th frame image and institute Shot segmentation point of the L+n frames image as the video is stated, and will be between the L-th frame image and the L+n frame images Picture frame as the picture frame except at least one camera lens picture, wherein the n be greater than or equal to 1.

In some embodiments of the invention, aforementioned schemes are based on, the difference computing unit is used for：Described in calculating successively The gray scale difference value of adjacent two field pictures in video；Or the color histogram feature of adjacent two field pictures in the video is calculated successively Difference；Or the difference of the edge feature of adjacent two field pictures in the video is calculated successively.

In some embodiments of the invention, aforementioned schemes are based on, second generation unit is used for：From the video Extraction includes the picture frame of specified content as the key frame.

In some embodiments of the invention, aforementioned schemes are based on, the exhibition scheme selecting unit is used for：It obtains and passes through Trained prediction model；Feature based on the user and the content characteristic predict described information by the prediction model Pass through the probability selected by the user when each candidate exhibition scheme displaying；By the highest candidate displaying of corresponding probability Scheme is as the intended display scheme.

In some embodiments of the invention, aforementioned schemes are based on, the first acquisition unit is used for：According to the user Feature, corresponding with the feature of user information, and the information conduct that will match to are matched from existing information bank Information to be recommended to the user.

In some embodiments of the invention, aforementioned schemes are based on, the information recommending apparatus further includes：Rationale for the recommendation Generation unit, for generating the reasons why recommending described information to the user；The recommendation unit is additionally operable to send out the reason Give the user.

One side according to the ... of the embodiment of the present invention provides a kind of computer-readable medium, is stored thereon with computer Program realizes the information recommendation method as described in above-described embodiment when the computer program is executed by processor.

One side according to the ... of the embodiment of the present invention provides a kind of electronic equipment, including：One or more processors； Storage device, for storing one or more programs, when one or more of programs are held by one or more of processors When row so that one or more of processors realize the information recommendation method as described in above-described embodiment.

In the technical solution that some embodiments of the present invention are provided, by obtaining the information to be recommended to user, and Used intended display scheme when directional user's recommendation information true according to the feature of user, enabling based on the spy with user It levies the exhibition scheme to match and to realize the personalized displaying of recommendation information to user's recommendation information, and then can improve and push away The clicking rate for the information recommended.

It should be understood that above general description and following detailed description is only exemplary and explanatory, not It can the limitation present invention.

Description of the drawings

The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the present invention Example, and be used to explain the principle of the present invention together with specification.It should be evident that the accompanying drawings in the following description is only the present invention Some embodiments for those of ordinary skill in the art without creative efforts, can also basis These attached drawings obtain other attached drawings.In the accompanying drawings：

Fig. 1 shows the exemplary system of the information recommendation method or information recommending apparatus that can apply the embodiment of the present invention The schematic diagram of framework；

Fig. 2 shows the structural schematic diagrams of the computer system for the electronic equipment for being suitable for being used for realizing the embodiment of the present invention；

Fig. 3 diagrammatically illustrates the flow chart of information recommendation method according to an embodiment of the invention；

Fig. 4 shows the display effect schematic diagram of recommendation information according to an embodiment of the invention；

Fig. 5 diagrammatically illustrates the flow chart of information recommendation method according to another embodiment of the invention；

Fig. 6 diagrammatically illustrates a kind of flow chart for realizing process of step S520 shown in Fig. 5；

Fig. 7 diagrammatically illustrates the flow chart according to an embodiment of the invention for generating candidate exhibition scheme；

Fig. 8 diagrammatically illustrates the flow chart of the candidate exhibition scheme of generation according to another embodiment of the invention；

Fig. 9 diagrammatically illustrates the flow chart according to an embodiment of the invention that key frame is extracted from video；

Figure 10 diagrammatically illustrate it is according to an embodiment of the invention by video slicing be at least one camera lens picture Flow chart；

Figure 11 diagrammatically illustrates a kind of flow chart for realizing process of step S630 shown in Fig. 6；

Figure 12 diagrammatically illustrates the flow chart of information recommendation method according to still another embodiment of the invention；

Figure 13 diagrammatically illustrates according to an embodiment of the invention according to the personalized displaying side of recommendation determination The flow chart of case；

Figure 14 diagrammatically illustrates the pass in the scheme extraction video according to an embodiment of the invention based on cluster The flow chart of key frame；

Figure 15 is diagrammatically illustrated according to an embodiment of the invention determines whether user can click recommendation based on LR models The flow chart of content；

Figure 16 diagrammatically illustrates the block diagram of information recommending apparatus according to an embodiment of the invention；

Figure 17 diagrammatically illustrates the block diagram of information recommending apparatus according to another embodiment of the invention；

Figure 18 diagrammatically illustrates the block diagram of processing unit according to an embodiment of the invention；

Figure 19 diagrammatically illustrates the block diagram of exhibition scheme generation unit according to an embodiment of the invention；

Figure 20 diagrammatically illustrates the block diagram of exhibition scheme generation unit according to another embodiment of the invention；

Figure 21 diagrammatically illustrates the block diagram of picture generation unit according to an embodiment of the invention；

Figure 22 diagrammatically illustrates the block diagram of the second generation unit according to an embodiment of the invention；

Figure 23 diagrammatically illustrates the block diagram of the second generation unit according to another embodiment of the invention；

Figure 24 diagrammatically illustrates the block diagram of cutting unit according to an embodiment of the invention.

Specific implementation mode

Example embodiment is described more fully with reference to the drawings.However, example embodiment can be with a variety of shapes Formula is implemented, and is not understood as limited to example set forth herein；On the contrary, thesing embodiments are provided so that the present invention will more Fully and completely, and by the design of example embodiment comprehensively it is communicated to those skilled in the art.

In addition, described feature, structure or characteristic can be incorporated in one or more implementations in any suitable manner In example.In the following description, many details are provided to fully understand the embodiment of the present invention to provide.However, It will be appreciated by persons skilled in the art that technical scheme of the present invention can be put into practice without one or more in specific detail, Or other methods, constituent element, device, step may be used etc..In other cases, it is not shown in detail or describes known side Method, device, realization or operation are to avoid fuzzy each aspect of the present invention.

Block diagram shown in attached drawing is only functional entity, not necessarily must be corresponding with physically separate entity. I.e., it is possible to realize these functional entitys using software form, or realized in one or more hardware modules or integrated circuit These functional entitys, or these functional entitys are realized in heterogeneous networks and/or processor device and/or microcontroller device.

Flow chart shown in attached drawing is merely illustrative, it is not necessary to including all content and operation/step, It is not required to execute by described sequence.For example, some operation/steps can also decompose, and some operation/steps can close And or part merge, therefore the sequence actually executed is possible to be changed according to actual conditions.

Fig. 1 shows the exemplary system of the information recommendation method or information recommending apparatus that can apply the embodiment of the present invention The schematic diagram of framework 100.

As shown in Figure 1, system architecture 100 may include one or more, the network in terminal device 101,102,103 104 and server 105.Network 104 is to the offer communication link between terminal device 101,102,103 and server 105 Medium.Network 104 may include various connection types, such as wired, wireless communication link or fiber optic cables etc..

It should be understood that the number of the terminal device, network and server in Fig. 1 is only schematical.According to realization need It wants, can have any number of terminal device, network and server.For example server 105 can be multiple server compositions Server cluster etc..

User can be interacted by network 104 with server 105 with using terminal equipment 101,102,103, to receive or send out Send message etc..Terminal device 101,102,103 can be the various electronic equipments for having display screen, including but not limited to intelligent hand Machine, tablet computer, portable computer and desktop computer etc..

Server 105 can be to provide the server of various services.Such as server 105 can get it is to be recommended to After the information of user, used exhibition scheme when being determined to user's recommendation information according to the feature of user, and then be based on Determining exhibition scheme will send information to the terminal device 103 (can also be terminal device 101 or 102) of user, due to letter The exhibition scheme of breath is matched with the feature of user, it is achieved that the personalized displaying of recommendation information, is conducive to improve The clicking rate of the information of recommendation.

Meanwhile server 105 can determine the information to be recommended to user according to the feature of user, can ensure in this way Information recommended to the user and the user feature of itself match, and improve the accuracy of information recommendation.

It should be noted that the information recommendation method that the embodiment of the present invention is provided generally is executed by server 105, accordingly Ground, information recommending apparatus are generally positioned in server 105.

Fig. 2 shows the structural schematic diagrams of the computer system for the electronic equipment for being suitable for being used for realizing the embodiment of the present invention.

It should be noted that Fig. 2 shows the computer system 200 of electronic equipment be only an example, should not be to this hair The function and use scope of bright embodiment bring any restrictions.

As shown in Fig. 2, computer system 200 includes central processing unit (CPU) 201, it can be read-only according to being stored in Program in memory (ROM) 202 or be loaded into the program in random access storage device (RAM) 203 from storage section 208 and Execute various actions appropriate and processing.In RAM 203, it is also stored with various programs and data needed for system operatio.CPU 201, ROM202 and RAM 203 are connected with each other by bus 204.Input/output (I/O) interface 205 is also connected to bus 204。

It is connected to I/O interfaces 205 with lower component：Importation 206 including keyboard, mouse etc.；It is penetrated including such as cathode The output par, c 207 of spool (CRT), liquid crystal display (LCD) etc. and loud speaker etc.；Storage section 208 including hard disk etc.； And the communications portion 209 of the network interface card including LAN card, modem etc..Communications portion 209 via such as because The network of spy's net executes communication process.Driver 210 is also according to needing to be connected to I/O interfaces 205.Detachable media 211, such as Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on driver 210, as needed in order to be read from thereon Computer program be mounted into storage section 208 as needed.

Particularly, according to an embodiment of the invention, it may be implemented as computer below with reference to the process of flow chart description Software program.For example, the embodiment of the present invention includes a kind of computer program product comprising be carried on computer-readable medium On computer program, which includes the program code for method shown in execution flow chart.In such reality It applies in example, which can be downloaded and installed by communications portion 209 from network, and/or from detachable media 211 are mounted.When the computer program is executed by central processing unit (CPU) 201, executes and limited in the system of the application Various functions.

It should be noted that computer-readable medium shown in the present invention can be computer-readable signal media or meter Calculation machine readable storage medium storing program for executing either the two arbitrarily combines.Computer readable storage medium for example can be --- but not Be limited to --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor system, device or device, or arbitrary above combination.Meter The more specific example of calculation machine readable storage medium storing program for executing can include but is not limited to：Electrical connection with one or more conducting wires, just It takes formula computer disk, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type and may be programmed read-only storage Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device, Or above-mentioned any appropriate combination.In the present invention, can be any include computer readable storage medium or storage journey The tangible medium of sequence, the program can be commanded the either device use or in connection of execution system, device.And at this In invention, computer-readable signal media may include in a base band or as the data-signal that a carrier wave part is propagated, Wherein carry computer-readable program code.Diversified forms may be used in the data-signal of this propagation, including but unlimited In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for By instruction execution system, device either device use or program in connection.Include on computer-readable medium Program code can transmit with any suitable medium, including but not limited to：Wirelessly, electric wire, optical cable, RF etc. or above-mentioned Any appropriate combination.

Flow chart in attached drawing and block diagram, it is illustrated that according to the system of various embodiments of the invention, method and computer journey The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation A part for a part for one module, program segment, or code of table, above-mentioned module, program segment, or code includes one or more Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical On can be basically executed in parallel, they can also be executed in the opposite order sometimes, this is depended on the functions involved.Also it wants It is noted that the combination of each box in block diagram or flow chart and the box in block diagram or flow chart, can use and execute rule The dedicated hardware based systems of fixed functions or operations is realized, or can use the group of specialized hardware and computer instruction It closes to realize.

Being described in unit involved in the embodiment of the present invention can be realized by way of software, can also be by hard The mode of part realizes that described unit can also be arranged in the processor.Wherein, the title of these units is in certain situation Under do not constitute restriction to the unit itself.

As on the other hand, present invention also provides a kind of computer-readable medium, which can be Included in electronic equipment described in above-described embodiment；Can also be individualism, and without be incorporated the electronic equipment in. Above computer readable medium carries one or more program, when the electronics is set by one for said one or multiple programs When standby execution so that the electronic equipment realizes the method as described in following embodiments.For example, the electronic equipment can be real Now such as Fig. 3, Fig. 5 to each step shown in figure 15.

The realization details of the technical solution of the embodiment of the present invention is described in detail below：

In one embodiment of the invention, it when to user's recommendation information, according to the personal characteristics of user and can push away The feature for recommending content is matched, and the information to find most suitable user pushes.Specifically as shown in figure 3, according to the present invention One embodiment information recommendation method (information recommendation method be suitable for previous embodiment described in electronic equipment), packet Include following steps：

Step S310 extracts user characteristics.

In one embodiment of the invention, user characteristics are according to the social property of user, living habit and consumption row For the user model that equal informations go out, for example, can be based on user age, residence, gender, work, school, interest love Good, income etc. determines user characteristics.

Step S320 extracts content characteristic.

In one embodiment of the invention, classification that can be based on content, the keyword of content, the corresponding trip of content Play, content pouplarity etc. carry out the feature of content construction.

It should be noted that the embodiment of the present invention does not make the execution sequencing of step S310 and step S320 It is specific to limit, when actually executing, step S310 can be both first carried out, then execute step S320；Or step can also be first carried out Rapid S320, then execute step S310；Step S310 and step S320 can also be performed simultaneously.

Step S330, user characteristics and content characteristic combined crosswise.

In one embodiment of the invention, can establish a recommended models, and according to the feature of sample data come pair The recommended models are trained, then by the content extracted in the user characteristics extracted in step S310 and step S320 spy Recommended models after sign input training, to determine recommendation results according to the output result of the recommended models after training.

Step S340 generates recommendation results.

In one embodiment of the invention, the output of recommended models is the result is that a probability value can in specific choice Using the maximum content of select probability value as result recommended to the user.Since content recommended to the user is the feature with user Match, therefore can ensure that content recommended to the user is more in line with the demand of user.

Step S350, recommendation displaying.

In one embodiment of the invention, as shown in figure 4, when to user's recommendation, content is usually shown Title, abstract and thumbnail, specifically as shown in 402 in Fig. 4.Therefore, when to user's recommendation, the content not only recommended It affects user and whether can click and check, and the content that recommendation how is presented can also influence the decision of user, and then affect The clicking rate of recommendation.

Based on this, the embodiment provides following prioritization schemes：

Fig. 5 diagrammatically illustrates the flow chart of information recommendation method according to another embodiment of the invention, the information Recommendation method is suitable for the electronic equipment described in previous embodiment.Referring to Figure 5, which includes at least step Rapid S510 is described in detail as follows to step S530：

In step S510, the information to be recommended to user is obtained.

In one embodiment of the invention, the information to be recommended to user can be advertisement, game, article, video etc.. Wherein, the information to be recommended to user can be identical content for all users, can also be different content.

For example, information corresponding with the feature of user can be matched from existing information bank according to the feature of user, And the information that will match to is as the information to be recommended to the user.It is corresponding with the feature of user being matched from information bank Information when can refer to above-mentioned process shown in Fig. 3, i.e., according to the feature of each information in the feature of user and information bank come Information recommended to the user is determined, to meet the individual demand of different user, it is ensured that the information recommended to each user can Match with the unique characteristics of user, and then improves the accuracy for the information recommended.

It is used when determining to user recommendation described information according to the feature of the user in step S520 Intended display scheme.

In an embodiment of the present invention, used target when directional user's recommendation information true by feature according to user Exhibition scheme, enabling the exhibition scheme that is matched based on the feature with user to user's recommendation information, realizes recommendation The personalized displaying of information, and then the clicking rate of the information of recommendation can be improved.For the realization details of step S520 will with It is described in detail in lower content.

Shown in Fig. 5, in step S530, the letter is recommended to the user based on the intended display scheme Breath.

It in one embodiment of the invention, can be according to mode shown in Fig. 4 come to user's recommendation information, you can with By forms such as thumbnail, title, abstracts come to user's recommendation information.It is different by that can be directed in an embodiment of the present invention User uses different exhibition schemes, and corresponding thumbnail, title and abstract may when such as same information is recommended to different user And differ, it is achieved that personalized displaying, is conducive to the clicking rate for improving the information recommended.For example, a same trip Play video, can emphasize the excellent degree of operation on title when recommending superior player, can on title when recommending leisure player To emphasize the interesting property of content；If it is the player of heroic A, then thumbnail can intercept the performances of heroic A in video, if It is the player of heroic B, then thumbnail will choose the picture that can protrude hero B, be watched so as to more attract user to click Content.

In one embodiment of the invention, the reasons why recommending described information to the user can also be generated, and will be given birth to At the reasons why be sent to the user.It in an embodiment of the present invention, should be reason may is that based on good friend's viewing, user's focus Equal generations.

The realization details of step S520 shown in Fig. 5 is described in detail below：

As shown in fig. 6, in one embodiment of the invention, according to the feature of user in step S520, determining to described The process of used intended display scheme may include step S610 to step S630 when user recommends described information, right below The realization details of each step is described in detail：

In step S610, multiple candidate exhibition schemes are generated according to the information to be recommended to user.

In one embodiment of the invention, may include in the information to be recommended to user content of text, picture and Video etc. illustrates how to generate above-mentioned candidate exhibition scheme below for content of text, picture and video respectively：

Candidate exhibition scheme is generated for content of text：

In one embodiment of the invention, with reference to shown in Fig. 7, the candidate exhibition of generation according to an embodiment of the invention Show the flow of scheme, including：

Step S710 extracts the content of text in the information to be recommended to user.

In one embodiment of the invention, it can be identified based on text recognition technique in the information to be recommended to user Including content of text, then extract the content of text recognized.

Step S720 generates the abstract and title of multiple candidates according to the content of text.

In one embodiment of the invention, the abstract of multiple candidates is generated according to the content of text in step S720 And title, including：For multiple dimensions of the content of text, abstract corresponding with each dimension in the multiple dimension is generated And title, to obtain the synopsis and title of the multiple candidate.

Specifically, in one embodiment of the invention, multiple dimensions of content of text may include information dimension (as dashed forward Go out the different role mentioned in information to be recommended), user's dimension (as prominent information to be recommended was checked by multiple good friends), propagate Dimension (temperature of such as prominent information to be recommended), keyword dimension (such as some hot words for including in prominent information to be recommended).

It should be noted that the process for the abstract and title for generating multiple candidates according to content of text can pass through machine The method of study automatically generates, for example establishes machine learning model, is carried out to machine learning model by training sample Training, and then the candidate abstract and title of content of text are exported based on the model after training.

It is multiple by being generated according to the content of text in the information to be recommended to user in the above embodiment of the present invention Candidate abstract and title, and according in the information to be recommended to user picture and/or video generate multiple candidate's pictures, Make it possible to automatically generate multiple candidate exhibition schemes, improves the formation efficiency of candidate exhibition scheme.

Candidate exhibition scheme is generated for picture and/or video：

In one embodiment of the invention, with reference to shown in Fig. 8, generation according to another embodiment of the invention is candidate The flow of exhibition scheme, including：

Step S810 extracts picture and/or video in the information to be recommended to user；

Step S820 generates multiple candidate pictures according to the picture and/or video.

It in one embodiment of the invention, can be using the picture extracted from the information to be recommended to user as described in Candidate picture；And/or at least one key frame is extracted from the video that the information to be recommended to user includes, at least one by described in A key frame is as the candidate picture.

In one embodiment of the invention, the picture frame comprising specified content can be extracted from video as crucial Frame, for example extraction includes the picture frame of a certain role as key frame from video.

In another embodiment of the present invention, can by include in video picture frame carry out clustering processing come To key frame, detailed process is as shown in figure 9, comprise the following processes：

Step S910, the picture frame for including by the video carry out clustering processing, obtain at least one class.

In an embodiment of the present invention, hierarchical clustering algorithm may be used in the algorithm for carrying out clustering processing, K-means is calculated Method, K-medoids algorithms etc..For example when using K-means algorithms, cluster process includes the following steps：

(1) arbitrarily select k object as initial cluster center in the picture frame for including from video；

(2) each picture frame is calculated at a distance from this k cluster centre, and according to minimum range again to each picture frame Affiliated class is divided；

(3) cluster centre of each cluster is recalculated；

(4) canonical measure function is calculated, when meeting certain condition, when such as function convergence, then algorithm terminates；If condition is not Satisfaction then returns to above-mentioned steps (2).

Step S920, the nearest picture frame of chosen distance cluster centre is as the key frame from each class.

It should be noted that due to including the picture of a plurality of lenses in video, and can add in the handoff procedure of camera lens picture Enter some spaces or temporal edit effect, such as be fade-in fade-out, therefore in one embodiment of the invention, in step Before S910, i.e., before the picture frame progress clustering processing for including by the video, can also include：It is by the video slicing At least one camera lens picture, and delete the picture frame except at least one camera lens picture.

In this embodiment, by being at least one camera lens picture by video slicing, and the figure except camera lens picture is deleted As frame, enabling delete some in space or temporal effect picture, picture of being such as fade-in fade-out, and then video can be reduced In effect picture influenced caused by clustering processing, to ensure to be capable of the reality of reflecting video by clustering determining key frame Matter content.

In one embodiment of the invention, as shown in Figure 10, according to an embodiment of the invention by video slicing It is elaborated as follows for the flow of at least one camera lens picture, including step S101 to step S103, each step：

In step S101, the difference of adjacent two field pictures in the video is calculated successively.

In one embodiment of the invention, the difference of adjacent two field pictures in the video is calculated in step S101 successively It can be the gray scale difference value for calculating adjacent two field pictures in video, can also be the color histogram spy for calculating adjacent two field pictures The difference of sign can also be the difference for the edge feature for calculating adjacent two field pictures.

In step s 102, according to the difference of adjacent two field pictures in the video, determine that the camera lens in the video is cut Branch.

In one embodiment of the invention, determine that the process of the point of the shot segmentation in video may include：If described regard The difference of L-th frame image and L-1 frame images in frequency is greater than or equal to first threshold, then by the L-1 frames image and institute State a shot segmentation point as the video between L-th frame image, wherein the L is greater than or equal to 2.

In this embodiment, when the difference of adjacent two field pictures is larger, illustrate that camera lens is switched, therefore can incite somebody to action A shot segmentation point is used as between L-1 frames image and L-th frame image.

In one embodiment of the invention, determining the process of the point of the shot segmentation in video can also include：If described The difference of L-th frame image and the L-1 frame images is less than the first threshold, and is greater than or equal to second threshold, then from institute The L-th frame image stated in video starts, the difference for the adjacent two field pictures that add up successively；When the L-th frame image with it is described The difference summation of adjacent two field pictures is greater than or equal to the first threshold, and the L between L+n frame images in video When the difference of L+n-1 frame images in+n frames image and the video is less than the second threshold, by the L-th frame image and Shot segmentation point of the L+n frames image as the video, and by the L-th frame image and the L+n frames image it Between picture frame as the picture frame except at least one camera lens picture, wherein the n be greater than or equal to 1.

In this embodiment, when the difference of L-th frame image and L-1 frame images is less than first threshold, and it is greater than or equal to When second threshold, illustrate from L-th frame be initially gradual shot switch starting position；When L-th frame image and L+n frames image it Between the difference summations of adjacent two field pictures be greater than or equal to first threshold, and the difference of L+n frames image and L+n-1 frame images When less than second threshold, illustrate that L+n frame images are the end positions of gradual shot switching, thus can by L-th frame image and Shot segmentation point of the L+n frames image as video, and using the picture frame between L-th frame image and L+n frame images as institute State the picture frame except at least one camera lens picture.

It is extremely by the video slicing based on determining shot segmentation point in step s 103 shown in Figure 10 A few camera lens picture.

In an embodiment of the present invention, by video slicing be at least one camera lens picture after, can to obtain to The picture frame for including in a few camera lens picture carries out clustering processing, to get the key frame in video based on cluster result.

Shown in Fig. 6, in step S620, the content characteristic of each candidate exhibition scheme is obtained.

In one embodiment of the invention, candidate exhibition scheme may include title, abstract and thumbnail, candidate's displaying The content characteristic of scheme is title feature, abstract feature and picture feature etc..

Shown in Fig. 6, in step S630, according to the content characteristic, from the multiple candidate exhibition scheme It is middle to select the candidate exhibition scheme to match with the feature of the user as the intended display scheme.

In one embodiment of the invention, referring to Fig.1 shown in 1, step S630 may include：

Step S111 obtains trained prediction model.

In one embodiment of the invention, prediction model can be trained according to sample data, after training The output of prediction model is the result is that a probability value.

Step S112, the feature based on the user and the content characteristic predict the letter by the prediction model The probability that breath is selected when being shown by each candidate exhibition scheme by the user；

Step S113, using the highest candidate exhibition scheme of corresponding probability as the intended display scheme.

In this embodiment, since the intended display scheme of selection is the highest displaying side of probability that prediction model predicts Case, therefore, to user's recommendation information, recommendation information can be realized based on the exhibition scheme that the feature with user matches Personalization displaying, and then the clicking rate of information recommended to the user can be improved.

The technical solution of the embodiment of the present invention and realization details are elaborated above, below to recommend to user Include text, picture and video content for, to a concrete application field of the information recommendation scheme of the embodiment of the present invention Scape illustrates.

In the concrete application scene of the present invention, as shown in figure 12, letter according to still another embodiment of the invention Suggested design is ceased, is mainly comprised the following processes：

Content characteristic is input to recommended models by step S121.

It in one embodiment of the invention, can will be each in content library (database for storing content to be recommended) By Feature Engineering, (Feature Engineering is algorithm can be made to reach optimum performance to create using the relevant knowledge of data fields to content Feature process) get content characteristic.

User's portrait is input to recommended models by step S122.

In one embodiment of the invention, the feature of user can be got by Feature Engineering, and then is built and used It draws a portrait at family.

It should be noted that the embodiment of the present invention does not make the execution sequencing of step S121 and step S122 It is specific to limit, when actually executing, step S121 can be both first carried out, then execute step S122；Or step can also be first carried out Rapid S122, then execute step S121；Step S121 and step S122 can also be performed simultaneously.

In one embodiment of the invention, recommended models are that (sample data can be according to history number by sample data According to obtaining) machine learning model trained.

Step S123, recommended models draw a portrait according to the content characteristic of input and user and export recommendation.

Recommendation is input to displaying model by step S124.

User's portrait is input to displaying model by step S125.

It should be noted that the embodiment of the present invention does not make the execution sequencing of step S124 and step S125 It is specific to limit, when actually executing, step S124 can be both first carried out, then execute step S125；Or step can also be first carried out Rapid S125, then execute step S124；Step S124 and step S125 can also be performed simultaneously.

In one embodiment of the invention, displaying model is also that (sample data can be according to history by sample data Data obtain) machine learning model trained.

Step S126, displaying model draw a portrait according to the recommendation of input and user and export personalized displaying content, i.e., It is drawn a portrait based on user, the exhibition scheme of recommendation is determined by showing model, such as shown what thumbnail, show that is marked Topic and abstract, reason recommended to the user etc..

Flow shown in Figure 12 contains the process of determining recommendation and determines personalized displaying according to recommendation The process of scheme, wherein determine that the process of recommendation is referred to shown in Fig. 3, personalized displaying is determined according to recommendation The process of scheme is referred to shown in Figure 13, specifically comprise the following steps：

Step S131 gets user's portrait by Feature Engineering.

Recommendation is divided into text and picture/video by step S132.

Step S133, for text generation text snippet and title.

In one embodiment of the invention, multiple abstracts and title can be generated from different angles, using as exhibition Show the candidate of text.Wherein, different angles is the not ipsilateral for protruding text, such as an article, Ke Yitong Following dimension is crossed to generate synopsis：

A) information dimension：The different role mentioned in such as prominent article

B) user's dimension：Such as prominent article has been seen by multiple good friends

C) propagation dimension：The temperature of such as prominent article

D) keyword dimension：Some hot words of such as prominent article

It should be noted that in an embodiment of the present invention, the abstract and title of generation either by manually generated, It can also be automatically generated by the method for machine learning.

Step S134 generates picture abstract according to picture/video.The process includes mainly generating candidate picture for picture Process and the process of candidate picture is generated for video.

It in one embodiment of the invention, can be by these pictures if in recommendation including multiple pictures As the candidate picture of displaying, to show different pictures for different users.

In one embodiment of the invention, it for the video for including in recommendation, can be extracted from video multiple Key frame, using the multiple key frames extracted as the candidate picture of displaying.

In an embodiment of the present invention, in the key frame in extracting video, the method based on camera lens may be used, be based on The method of motion analysis, the method etc. based on cluster.Wherein, the method based on camera lens is that video is divided into a plurality of lenses picture, One or more key frames are chosen from each camera lens picture；Method based on motion analysis is to determine adjacent image frame in video Between amplitude of variation, if amplitude of variation is larger, as different classes, if amplitude of variation is smaller, be used as identical class, Then one or more key frames are chosen from each class.

Method described in detail below based on cluster：

Detailed process includes referring to Fig.1 mainly step S141 and step S142 shown in 4.

In step s 141, it is a plurality of lenses picture by Video segmentation.

In an embodiment of the present invention, by Video segmentation at a plurality of lenses picture, the boundary of video lens is mainly found, One camera lens picture is an event for describing Same Scene or continuous action, therefore when camera lens picture is changed It waits, the variation occurred between picture frame can be bigger, and the difference of the picture frame in the same camera lens picture can be smaller, because This can be by judging the difference of adjacent two field pictures to determine whether belonging to the same camera lens picture.

In one embodiment of the invention, the difference between two field pictures can by calculate gray scale difference method come It arrives, it is specific as shown in formula 1：

In equation 1, Dis indicates the difference between two field pictures；M indicates two field pictures I₁And I₂Pixel number；X and y tables Diagram is as I₁And I₂The position of corresponding pixel.

In other embodiments of the invention, the difference between two field pictures can also be by comparing the color of two field pictures The methods of histogram feature and edge feature obtain.

Since the transformation of camera lens includes：Mutation and gradual change.Mutation refer to directly be converted into from a camera lens in video it is next A camera lens；Gradual change refers to that some spaces or temporal edit effect is added in handoff procedure in camera lens, is such as fade-in fade-out Deng.Therefore when determining whether camera lens switches, dual threshold comparison method can used by comparing the difference between picture frame It determines, concrete scheme is as follows：

Step (1) calculates the frame of present frame (i.e. L-th frame) and former frame (i.e. L-1 frames) for the L-th frame in video Between difference Dis (I_L,I_L-1)；Then go to step (2).

Step (2), if current frame-to-frame differences (i.e. above-mentioned Dis (I_L,I_L-1)) it is more than threshold value T1, then it is assumed that the frame (i.e. the L frames) be lens mutation switch frame, then carry out camera lens cutting.If current frame-to-frame differences is less than threshold value T1, but is greater than threshold value T2, then it is assumed that the frame (i.e. L-th frame) is the starting position of gradual shot switching, goes to step (3)；If current frame-to-frame differences is less than Threshold value T2, then it is assumed that the frame (i.e. L-th frame) is not the position of shot segmentation, goes back to step (1).

Step (3) calculates the current frame-to-frame differences of each frame since L+1 frames, and the frame-to-frame differences for adjacent two frame that adds up. If calculated to L+n frames, cumulative frame-to-frame differences is more than threshold value T1, and current frame-to-frame differences (i.e. Dis (I_L+n,I_L+n-1)) small In threshold value T2, then it is assumed that the frame (i.e. L+n frames) is the end position of gradual shot switching, carries out camera lens cutting.If cumulative Frame-to-frame differences be less than threshold value T1, and current frame-to-frame differences (i.e. Dis (I_L+n,I_L+n-1)) again smaller than threshold value T2, then it is assumed that this is not one Secondary camera lens Gradual change goes to step (1) and continues to calculate, until by the picture frame in video, all traversal is completed.

In the above-described embodiments, by carrying out shot segmentation to video, enabling the gradual change image that will include in video Frame (picture frame i.e. between the starting position and end position of Gradual change) is deleted, to avoid subsequently when extracting key frame Lead to the key frame inaccuracy of extraction due to there is a problem of for gradual change picture frame.

In step S142, key frame is extracted from the lens image that segmentation obtains.

In an embodiment of the present invention, characteristics of image can all be extracted to each frame image in obtained a plurality of lenses (such as color characteristic, histogram feature, shape feature, motion feature etc.) is then based on K-Means methods all picture frames It is divided into K classes, a frame nearest from cluster centre is finally chosen in every one kind as such key frame.

It should be noted that when carrying out clustering processing, other clustering algorithms, such as hierarchical clustering can also be used to calculate Method, K-medoids algorithms etc..

Shown in Figure 13, determine that the process of personalized exhibition scheme can also include according to recommendation：

User's portrait is input to machine learning model by step S135a；Step S135b, by determining candidate abstract and mark Topic is input to machine learning model；Determining candidate picture is input to machine learning model by step S135c.Wherein, of the invention Embodiment considered critical is not carried out to the execution of step S135a, step S135b and step S135c sequence.

Step S136, machine learning model export personalization according to user's portrait, candidate abstract and title, candidate picture Recommendation exhibition scheme.

In one embodiment of the invention, it can be drawn a portrait to obtain user characteristics according to user, to candidate's abstract and title Text feature is extracted, the features such as color histogram are extracted to candidate picture, then various features are intersected, use LR models (Logistic Regression, logistic regression) prediction user clicks the probability of the recommendation.Wherein, abstract and title Text feature includes that (Term Frequency-Inverse Document Frequency, one kind being used for information based on TF-IDF Retrieval and the common weighting technique of data mining) bag of word features and doc2vec features etc..

The above process is referred to：User characteristics are input to LR models by step S152a；Step Text feature is input to LR models by rapid S152b；Picture feature is input to LR models by step S152c.Step S154, passes through LR models determine whether user can click the content of recommendation.Wherein, the embodiment of the present invention is not to step S152a, step S152b Considered critical is carried out with the execution sequence of step S152c.

Wherein, the output of LR models is the probability that user clicks the recommendation, therefore for each time of recommendation Scheme (being shown by which candidate abstract and title, by which candidate picture) is selected, a user can be obtained The probability whether clicked, and then can be using the highest exhibition scheme of select probability as the scheme of final displaying recommendation.

In one embodiment of the invention, before exporting the probability of each candidate scheme by above-mentioned LR models, It needs to be trained LR models.Wherein, the process being trained to LR models is as follows：

First, it is handled to obtain one group of training data D according to historical data：

D=(x¹,y¹),(x²,y²)…(x^N,y^N)

Wherein, x represents a M dimensional feature vector (feature for including user characteristics and recommendation)；Whether y represents user The content can be clicked, 1 indicates to click, and 0 indicates to click.

The purpose of model training is desirable to find a function f so that loss functionValue it is minimum. Under LR models, this function is defined as：

Based on above-mentioned function, the target of model training seeks to find one group of parameter θ that can minimize loss function, so New data are predicted using this group of parameter afterwards.

It should be noted that being illustrated so that LR models are predicted as an example in above-described embodiment, in its of the present invention In its embodiment, the machine learning algorithm except LR, such as neural network algorithm, random forests algorithm can also be chosen, specifically It can be chosen based on existing data and application scenarios.

Meanwhile above-described embodiment be by recommend to user include text, picture and video content for illustrate this One concrete application scene of the information recommendation scheme of inventive embodiments can also be right in the other application scene of the present invention Advertising information is recommended, is recommended etc. game, no matter which kind of information is recommended, in the application embodiment of the present invention After information recommendation scheme, it can be based on, to user's recommendation information, to realize with the exhibition scheme that the feature of user matches The personalized displaying of recommendation information, and then can improve the clicking rate of the information of recommendation.

The device of the invention embodiment introduced below can be used for executing the information recommendation side in the above embodiment of the present invention Method.For undisclosed details in apparatus of the present invention embodiment, the embodiment of above-mentioned information recommendation method of the invention is please referred to.

Figure 16 diagrammatically illustrates the block diagram of information recommending apparatus according to an embodiment of the invention.

Referring to Fig.1 shown in 6, information recommending apparatus 160 according to an embodiment of the invention, including：First obtains list Member 161, processing unit 162 and recommendation unit 163.

Wherein, first acquisition unit 161 is for obtaining the information to be recommended to user；Processing unit 162 is used for according to institute The feature of user is stated, used intended display scheme when determining to user recommendation described information；Recommendation unit 163 is used for Recommend described information to the user based on the intended display scheme.

Figure 17 diagrammatically illustrates the block diagram of information recommending apparatus according to another embodiment of the invention.

Referring to Fig.1 shown in 7, information recommending apparatus 170 according to another embodiment of the invention is believed shown in Figure 16 On the basis of ceasing recommendation apparatus, further include：Rationale for the recommendation generation unit 164.

Wherein, rationale for the recommendation generation unit 164 is for generating the reasons why recommending described information to the user；The recommendation Unit 163 is additionally operable to the reason being sent to the user.

In some embodiments of the invention, first acquisition unit 161 shown in Figure 16 and Figure 17 is used for：According to described The feature of user matches information corresponding with the feature of the user, and the information that will match to from existing information bank As the information to be recommended to the user.

Referring to Fig.1 shown in 8, in one embodiment of the invention, processing unit 162 shown in Figure 16 and Figure 17 wraps It includes：Exhibition scheme generation unit 1621, second acquisition unit 1622 and exhibition scheme selecting unit 1623.

Wherein, exhibition scheme generation unit 1621, for generating multiple candidate exhibition schemes according to described information；Second obtains Unit 1622 is taken, the content characteristic for obtaining each candidate exhibition scheme；Exhibition scheme selecting unit 1623 is used for root According to the content characteristic, the candidate displaying side that the feature of selection and the user match from the multiple candidate exhibition scheme Case is as the intended display scheme.

In some embodiments of the invention, aforementioned schemes are based on, the exhibition scheme selecting unit 1621 is used for：It obtains Trained prediction model；Feature based on the user and the content characteristic, described in prediction model prediction The probability that information is selected when being shown by each candidate exhibition scheme by the user；By the highest candidate of corresponding probability Exhibition scheme is as the intended display scheme.

Referring to Fig.1 shown in 9, in one embodiment of the invention, exhibition scheme generation unit 1621 includes：First extraction Unit 191 and abstract and title generation unit 192.

Wherein, the first extraction unit 191 is used to extract the content of text in described information；Abstract and title generation unit 192 for according to the content of text, generating the abstract and title of multiple candidates.

In some embodiments of the invention, aforementioned schemes are based on, abstract and title generation unit 192 are used for：For institute Multiple dimensions of content of text are stated, abstract corresponding with each dimension in the multiple dimension and title are generated, it is described to obtain The synopsis and title of multiple candidates.

With reference to shown in Figure 20, in another embodiment of the present invention, exhibition scheme generation unit 1621 includes：Second carries Take unit 193 and picture generation unit 194.

Wherein, the second extraction unit 193 is used to extract the picture and/or video in described information；Picture generation unit 194 For according to the picture and/or video, generating multiple candidate pictures.

It should be noted that in yet another embodiment of the present invention, on exhibition scheme generation unit 1621 may include The first extraction unit 191, abstract and title generation unit 192, the second extraction unit 193 and the picture generation unit 194 stated.

With reference to shown in Figure 21, in one embodiment of the invention, picture generation unit 194 includes：First generation unit 1941 and/or second generation unit 1942.

Wherein, the picture that the first generation unit 1941 is used to extract from described information is as the candidate picture；The Two generation units from the video for extracting at least one key frame, using at least one key frame as the candidate Picture.

With reference to shown in Figure 22, in one embodiment of the invention, the second generation unit 1942 includes：Cluster cell 221 With picture frame selecting unit 222.

Wherein, cluster cell 221 is used for the picture frame for including by the video and carries out clustering processing, obtains at least one Class；Picture frame selecting unit 222 be used for from each class the nearest picture frame of chosen distance cluster centre as the pass Key frame.

With reference to shown in Figure 23, in another embodiment of the present invention, the second generation unit 1942 is with cluster cell 221 and picture frame selecting unit 222 on the basis of, further include：Cutting unit 223, being used for will be described in the cluster cell 221 It is at least one camera lens picture by the video slicing, and described in deletion before the picture frame that video includes carries out clustering processing Picture frame except at least one camera lens picture.

With reference to shown in Figure 24, in one embodiment of the invention, cutting unit 223 includes：Difference computing unit 2231, Cut-off determination unit 2232 and execution unit 2233.

Wherein, difference computing unit 2231 is used to calculate the difference of adjacent two field pictures in the video successively；Cut-off Determination unit 2232 is used for the difference according to adjacent two field pictures in the video, determines the shot segmentation point in the video； Execution unit 2233 is used for based on determining shot segmentation point, is at least one camera lens picture by the video slicing.

In some embodiments of the invention, aforementioned schemes are based on, the cut-off determination unit 2232 is used for：Described The difference of L-th frame image and L-1 frame images in video be greater than or equal to first threshold when, will the L-1 frames image and As a shot segmentation point of the video between the L-th frame image, wherein the L is greater than or equal to 2.

In some embodiments of the invention, aforementioned schemes are based on, the cut-off determination unit 2232 is additionally operable to：Institute When stating L-th frame image with the difference of the L-1 frame images less than the first threshold, and being greater than or equal to second threshold, from The L-th frame image in the video starts, the difference for the adjacent two field pictures that add up successively；When the L-th frame image and institute The difference summation for stating adjacent two field pictures between the L+n frame images in video is greater than or equal to the first threshold, and described When L+n frames image and the difference of the L+n-1 frame images in the video are less than the second threshold, by the L-th frame figure The shot segmentation point of picture and the L+n frames image as the video, and by the L-th frame image and the L+n frame figures Picture frame as between is as the picture frame except at least one camera lens picture, wherein the n is greater than or equal to 1.

In some embodiments of the invention, aforementioned schemes are based on, the difference computing unit 2231 is used for：It calculates successively The gray scale difference value of adjacent two field pictures in the video；Or the color histogram of adjacent two field pictures in the video is calculated successively The difference of feature；Or the difference of the edge feature of adjacent two field pictures in the video is calculated successively.

In some embodiments of the invention, aforementioned schemes are based on, second generation unit 1942 is used for：It is regarded from described Extraction includes the picture frame of specified content as the key frame in frequency.

It should be noted that although being referred to several modules or list for acting the equipment executed in above-detailed Member, but this division is not enforceable.In fact, according to the embodiment of the present invention, it is above-described two or more The feature and function of module either unit can embody in a module or unit.Conversely, an above-described mould Either the feature and function of unit can be further divided into and embodied by multiple modules or unit block.

Through the above description of the embodiments, those skilled in the art is it can be readily appreciated that example described herein is implemented Mode can also be realized by software realization in such a way that software is in conjunction with necessary hardware.Therefore, according to the present invention The technical solution of embodiment can be expressed in the form of software products, the software product can be stored in one it is non-volatile Property storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.) in or network on, including some instructions are so that a calculating Equipment (can be personal computer, server, touch control terminal or network equipment etc.) is executed according to embodiment of the present invention Method.

Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to its of the present invention Its embodiment.This application is intended to cover the present invention any variations, uses, or adaptations, these modifications, purposes or Person's adaptive change follows the general principle of the present invention and includes undocumented common knowledge in the art of the invention Or conventional techniques.The description and examples are only to be considered as illustrative, and true scope and spirit of the invention are by following Claim is pointed out.

It should be understood that the invention is not limited in the precision architectures for being described above and being shown in the accompanying drawings, and And various modifications and changes may be made without departing from the scope thereof.The scope of the present invention is limited only by the attached claims.

Claims (19)

1. a kind of information recommendation method, which is characterized in that including：

Obtain the information to be recommended to user；

According to the feature of the user, used intended display scheme when determining to user recommendation described information；

Recommend described information to the user based on the intended display scheme.

2. information recommendation method according to claim 1, which is characterized in that according to the feature of the user, determine to institute Used intended display scheme when user's recommendation described information is stated, including：

Multiple candidate exhibition schemes are generated according to described information；

Obtain the content characteristic of each candidate exhibition scheme；

According to the content characteristic, the candidate that the feature of selection and the user match from the multiple candidate exhibition scheme Exhibition scheme is as the intended display scheme.

3. information recommendation method according to claim 2, which is characterized in that generate multiple candidate displayings according to described information Scheme, including：

Extract the content of text in described information；

According to the content of text, the abstract and title of multiple candidates are generated.

4. information recommendation method according to claim 3, which is characterized in that according to the content of text, generate multiple times The abstract and title of choosing, including：

For multiple dimensions of the content of text, abstract corresponding with each dimension in the multiple dimension and title are generated, To obtain the synopsis and title of the multiple candidate.

5. information recommendation method according to claim 2, which is characterized in that generate multiple candidate displayings according to described information Scheme, including：

Extract the picture and/or video in described information；

According to the picture and/or video, multiple candidate pictures are generated.

6. information recommendation method according to claim 5, which is characterized in that according to the picture and/or video, generate more A candidate's picture, including：

Using the picture extracted from described information as the candidate picture；And/or

At least one key frame is extracted from the video, using at least one key frame as the candidate picture.

7. information recommendation method according to claim 6, which is characterized in that extract at least one key from the video Frame, including：

The picture frame for including by the video carries out clustering processing, obtains at least one class；

The nearest picture frame of chosen distance cluster centre is as the key frame from each class.

8. information recommendation method according to claim 7, which is characterized in that carried out in the picture frame for including by the video Before clustering processing, further include：

It is at least one camera lens picture by the video slicing, and deletes the picture frame except at least one camera lens picture.

9. information recommendation method according to claim 8, which is characterized in that by the video slicing be at least one camera lens Picture, including：

The difference of adjacent two field pictures in the video is calculated successively；

According to the difference of adjacent two field pictures in the video, the shot segmentation point in the video is determined；

It is at least one camera lens picture by the video slicing based on determining shot segmentation point.

10. information recommendation method according to claim 9, which is characterized in that according to adjacent two field pictures in the video Difference, determine the shot segmentation point in the video, including：

If the difference of L-th frame image and L-1 frame images in the video is greater than or equal to first threshold, by the L- As a shot segmentation point of the video between 1 frame image and the L-th frame image, wherein the L is greater than or equal to 2.

11. information recommendation method according to claim 10, which is characterized in that according to adjacent two field pictures in the video Difference, determine the shot segmentation point in the video, further include：

If the difference of the L-th frame image and the L-1 frame images is less than the first threshold, and is greater than or equal to the second threshold Value, then since the L-th frame image in the video, the difference for the adjacent two field pictures that add up successively；

Be more than when the difference summation of adjacent two field pictures between the L+n frame images in the L-th frame image and the video or Equal to the first threshold, and the difference of the L+n frames image and the L+n-1 frame images in the video is less than described the When two threshold values, using the L-th frame image and the L+n frames image as the shot segmentation of video point, and by the L Picture frame between frame image and the L+n frame images is as the picture frame except at least one camera lens picture, wherein The n is greater than or equal to 1.

12. information recommendation method according to claim 9, which is characterized in that calculate adjacent two frame in the video successively The difference of image, including：

The gray scale difference value of adjacent two field pictures in the video is calculated successively；Or

The difference of the color histogram feature of adjacent two field pictures in the video is calculated successively；Or

The difference of the edge feature of adjacent two field pictures in the video is calculated successively.

13. information recommendation method according to claim 6, which is characterized in that extract at least one pass from the video Key frame, including：

Extraction includes the picture frame of specified content as the key frame from the video.

14. information recommendation method according to claim 2, which is characterized in that according to the content characteristic, from the multiple Select the candidate exhibition scheme to match with the feature of the user as the intended display scheme, packet in candidate exhibition scheme It includes：

Obtain trained prediction model；

It is each described to predict that described information passes through by the prediction model for feature based on the user and the content characteristic The probability selected by the user when candidate exhibition scheme displaying；

Using the highest candidate exhibition scheme of corresponding probability as the intended display scheme.

15. the information recommendation method according to any one of claim 1 to 14, which is characterized in that obtain to be recommended to use The information at family, including：

According to the feature of the user, information corresponding with the feature of the user is matched from existing information bank, and will The information being matched to is as the information to be recommended to the user.

16. the information recommendation method according to any one of claim 1 to 14, which is characterized in that further include：

Generate the reasons why recommending described information to the user；

The reason is sent to the user.

17. a kind of information recommending apparatus, which is characterized in that including：

First acquisition unit, for obtaining the information to be recommended to user；

Processing unit, for the feature according to the user, used target when determining to user recommendation described information Exhibition scheme；

Recommendation unit, for recommending described information to the user based on the intended display scheme.

18. a kind of computer-readable medium, is stored thereon with computer program, which is characterized in that the computer program is located Manage the information recommendation method realized when device executes as described in any one of claim 1 to 16.

19. a kind of electronic equipment, which is characterized in that including：

One or more processors；

Storage device, for storing one or more programs, when one or more of programs are by one or more of processing When device executes so that one or more of processors realize the information recommendation side as described in any one of claim 1 to 16 Method.